A data-driven wind turbine equipment fault prediction method and system

Abstract

The present invention relates to the field of data processing technology, and specifically to a data-driven method and system for predicting wind turbine equipment faults. The present invention clusters collected real-time operation data through a DBSCAN algorithm, removes outlier data therein, obtains real operation data of wind turbine equipment, and effectively avoids the influence of outlier data. The real operation data of the wind turbine equipment is normalized through a maximum and minimum normalization method to reduce the amount of calculation for subsequent operations. A bidirectional search algorithm is used to search for wind turbine equipment corresponding to abnormal operation data of wind turbine equipment and generate abnormal wind turbine equipment text data through data identification, thereby improving the speed and efficiency of the search process. Wind turbine equipment fault prediction result data is constructed and wind turbine equipment fault warning operations are performed to realize real-time warning of wind turbine equipment.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data-driven method and system for predicting wind turbine equipment failures. Background Art

[0002] A wind turbine is a mechanical device that converts wind energy into mechanical energy, transmits it to a generator, and then converts it into electrical energy through the generator.

[0003] Since the working environment of wind turbines is often harsh, various operating equipment in the units are easily affected by the natural environment and prone to failure. Therefore, real-time monitoring and fault prediction of wind turbine equipment become increasingly important.

[0004] The Chinese invention patent with publication number CN116861219B introduces a method for diagnosing wind turbine pitch control faults. The method collects the vibration signal of the wind turbine pitch control system to be tested, filters the vibration signal, removes electrical noise, and extracts the vibration characteristics and peak characteristics of the filtered signal to obtain the feature vector to be tested. The feature vector to be tested is compared with the standard feature vector to obtain the abnormality. When the abnormality is greater than the abnormal threshold, there is a fault in the wind turbine pitch control system. The acquired abnormal threshold has not been precisely calculated, and the accuracy of the fault diagnosis result cannot be guaranteed. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In order to solve the deficiencies in the background technology, the present invention provides a data-driven wind turbine equipment fault prediction method and system to achieve intelligent detection and real-time warning of wind turbine equipment faults.

[0007] (II) Technical solution

[0008] A data-driven wind turbine equipment fault prediction method comprises the following steps:

[0009] S1. Collect real-time operation data of wind turbine equipment through equipment monitoring devices;

[0010] S2. Using a clustering algorithm to clean up the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment;

[0011] S3, performing numerical processing on the actual operation data of the wind turbine generator set equipment to obtain standard operation data of the wind turbine generator set equipment;

[0012] S4. Setting a normal operating data interval set of wind turbine equipment through an intelligent optimization algorithm;

[0013] S5, numerically matching the standard operation data of the wind turbine generator set equipment with the corresponding normal operation data interval of the wind turbine generator set equipment, determining whether the wind turbine generator set equipment fails according to the matching result, and generating abnormal operation data of the wind turbine generator set equipment when the wind turbine generator set equipment fails;

[0014] S6. Generate corresponding wind turbine equipment text data according to the abnormal operation data of the wind turbine equipment;

[0015] S7. Construct wind turbine equipment failure prediction result data and perform wind turbine equipment failure warning operations.

[0016] The present invention clusters the collected real-time operation data through the DBSCAN algorithm, removes the outlier data therein, obtains the real operation data of the wind turbine equipment, and effectively avoids the influence of the outlier data; normalizes the real operation data of the wind turbine equipment through the maximum and minimum standardization method, and then uses the mean formula to perform average numerical measurement processing on the normalized data to obtain the standard operation data of the wind turbine equipment, thereby ensuring the accuracy of the acquired data and reducing the calculation amount of subsequent operations; sets the normal operation data interval of the wind turbine equipment through the artificial hummingbird optimization algorithm, thereby improving the accuracy of the setting result; and sets the standard operation data of the wind turbine equipment The quasi-operation data is numerically matched with the set normal operation data interval of the wind turbine equipment, and whether the wind turbine is faulty is determined based on the matching result. When the wind turbine equipment fails, the abnormal operation data of the wind turbine equipment is generated to ensure the accuracy of the judgment result and realize the intelligent detection of wind turbine equipment failure; a two-way search algorithm is used to search for the wind turbine equipment corresponding to the abnormal operation data of the wind turbine equipment and generate the abnormal equipment text data of the wind turbine after data identification, which improves the speed and efficiency of the search process; the wind turbine equipment failure prediction result data is constructed and the wind turbine equipment failure warning operation is performed to realize the real-time warning of the wind turbine equipment.

[0017] Preferably, the specific steps of collecting real-time operation data of wind turbine equipment through the equipment monitoring device are as follows:

[0018] S11. Set wind turbine equipment set ,in, Indicates Wind turbine operating equipment, Indicates the total number of wind turbine operating equipment;

[0019] S12: Set an operation data collection cycle, divide the operation data collection cycle evenly, and obtain an operation data collection time point set. ,in, Indicates The operation data collection time point, Indicates the total number of time points for running data collection;

[0020] S13, collecting the wind turbine equipment real-time operation data of each wind turbine equipment at each operation data collection time point through the equipment monitoring device, and obtaining the wind turbine equipment real-time operation data matrix as follows:

[0021] ,

[0022] in, Indicates that the equipment monitoring device is The first operation data collection time point is collected Real-time operating data of wind turbine operating equipment.

[0023] Preferably, the clustering algorithm is used to clean the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment in the following specific steps:

[0024] S21, using DBSCAN algorithm to analyze the real-time operation data matrix of wind turbine equipment The real-time operation data of each wind turbine operating equipment are clustered separately, and the real-time operation data matrix of wind turbine equipment is eliminated according to the clustering results. The outlier data in the real-time operation data of each wind turbine operating equipment is obtained to obtain the real operation data matrix of the wind turbine equipment. as follows:

[0025] ,

[0026] in, Indicates The first wind turbine operating equipment The real operating data of wind turbine equipment, It represents the total number of real operation data of wind turbine equipment in the real-time operation data of wind turbine equipment collected by the equipment monitoring device; the clustering process is as follows:

[0027] S211, set the neighborhood radius to And the data volume threshold in the core point neighborhood is ;

[0028] S212, randomly select a wind turbine operating device, and generate a real-time operating data matrix of the wind turbine device All real-time operating data of the wind turbine operating equipment described in the above are taken as a data set;

[0029] S213: Select any data in the data set, calculate the Euclidean distance between the data set and other data in the data set, and calculate the Euclidean distance less than the neighborhood radius. The total number of ;

[0030] S214, Comparison and If the numerical value of , then the selected data is determined to be a core point; otherwise, the selected data is determined to be an outlier;

[0031] S215, repeating the operations from S213 to S214 until all the data in the data set are traversed, and then eliminating outlier points in the data set that are not within the neighborhood radius of any core point.

[0032] The present invention clusters the collected real-time operation data through the DBSCAN algorithm, removes outlier data therein, obtains the real operation data of the wind turbine equipment, and effectively avoids the influence of outlier data.

[0033] Preferably, the specific steps of numerically processing the real operation data of the wind turbine equipment to obtain the standard operation data of the wind turbine equipment are as follows:

[0034] S31, the actual operation data matrix of the wind turbine equipment is normalized by the maximum and minimum standardization method. The real operation data of the wind turbine equipment in the normalization process is processed to obtain the normalized operation data matrix of the wind turbine equipment as follows:

[0035] ,

[0036] in, Indicates the actual operation data matrix of wind turbine equipment The first wind turbine operating equipment The normalized operating data of wind turbine equipment is obtained after normalization of the actual operating data of each wind turbine equipment; the calculation formula is as follows:

[0037] ,

[0038] in, and They represent the real operation data matrix of wind turbine equipment. The maximum and minimum values ​​of the actual operation data of all wind turbine equipment in the wind turbine operation equipment;

[0039] S32, using the mean formula to normalize the operation data matrix of the wind turbine equipment The normalized operation data of the wind turbine equipment in the system is processed by average numerical measurement to obtain the standard operation data set of the wind turbine equipment. ,in, Indicates The wind turbine equipment standard operating data of each wind turbine operating equipment, and , Indicates the total number of wind turbine operating devices.

[0040] The actual operation data of the wind turbine equipment is normalized by the maximum and minimum standardization method, and then the normalized data is averaged and measured by the mean formula to obtain the standard operation data of the wind turbine equipment, thereby ensuring the accuracy of the acquired data and reducing the amount of calculation for subsequent operations.

[0041] Preferably, the specific steps of setting the normal operation data interval set of the wind turbine equipment by the intelligent optimization algorithm are as follows:

[0042] S41. Build a hummingbird population and set the hummingbird population size to , the current number of iterations is The maximum number of iterations is And the dimension of the running data interval search space is ;

[0043] Randomly select a wind turbine device and generate a random value between [0,1] operating data intervals as normal operating data interval search samples for the wind turbine device, each operating data interval is used as a food source, and each hummingbird in the hummingbird population is assigned to a food source;

[0044] Randomly generate the initial position set of the hummingbird population in the running data interval search space ,in, Indicates the first The initial positions of individual hummingbirds;

[0045] S42, calculating the fitness function value of the food source where each hummingbird individual in the hummingbird population is located; the fitness function formula is as follows:

[0046] ,

[0047] in, Indicates the first The fitness function value of the food source where the hummingbird individual is located, Indicates the first The operating data interval corresponding to the food source where the hummingbird individuals are located meets the normal working capacity of the wind turbine equipment. Indicates the correction value;

[0048] S43. Each hummingbird in the hummingbird population will randomly select any one of the flight modes of axial flight, omnidirectional flight and diagonal flight to conduct guided foraging or territorial foraging in the search space of the running data interval;

[0049] S431. When a hummingbird individual is guided to forage in the search space of the running data interval, the hummingbird individual will select the food source where other hummingbird individuals are located as the target food source according to probability; the probability of each hummingbird individual being selected is as follows:

[0050] ,

[0051] in, Indicates The probability that the food source where the hummingbird individual is located is selected as the target food source;

[0052] After a hummingbird has selected a target food source, it will update its position according to the location of the target food source. The position update formula is as follows:

[0053] ,

[0054] in, Indicates the first The hummingbird individuals The location of the candidate food source selected after the location update during the iteration; Indicates the first The hummingbird individuals The position during the iteration; Indicates the first The hummingbird individuals The location of the target food source selected during the iteration; represents the guiding factor, which follows the normal distribution of (0,1); represents the flight mode chosen by individual hummingbirds;

[0055] S432. When a hummingbird individual is foraging in a territory in the search space of the running data interval, the hummingbird individual will search for a new food source near the food source to update its position; the position update formula is as follows:

[0056] ,

[0057] in, represents the regional factor, which follows the normal distribution of (0,1);

[0058] S44, calculating the fitness function value of the candidate food source selected by each hummingbird individual in the hummingbird population, if the fitness function value of the candidate food source selected by the hummingbird individual is greater than the fitness function value of the original food source, then the candidate food source is used to replace the original food source; otherwise, the original food source is retained;

[0059] S45, setting a migration threshold. If the current number of iterations is greater than or equal to the migration threshold, the hummingbird individual at the food source with the smallest fitness function value will continue to migrate for food, abandon the current food source, and randomly search for a new food source in the running data interval search space, and calculate the fitness function value of the food source.

[0060] S46, arranging the food sources of all hummingbird individuals according to the fitness function values, and selecting the food source with the highest fitness function value as the current optimal solution;

[0061] S47, determine the current number of iterations Is it less than the maximum number of iterations? , if the current number of iterations Less than the maximum number of iterations , then the current number of iterations Add 1 and return to S43; otherwise, take the current optimal solution as the global optimal solution, and output the operating data interval corresponding to the global optimal solution as the normal operating data interval of the selected wind turbine equipment;

[0062] S48, repeat the operations from S41 to S47 to generate a normal operating data interval set of the wind turbine equipment. ,in, Indicates The normal operating data range of each wind turbine equipment.

[0063] Compared with analyzing and processing the collected historical data to generate the normal operating data interval of the wind turbine equipment, the present invention sets the normal operating data interval of the wind turbine equipment through the artificial hummingbird optimization algorithm, which improves the accuracy of the setting results, speeds up the optimization process, and effectively avoids the optimization results from falling into the local optimum.

[0064] Preferably, the standard operation data of the wind turbine equipment is numerically matched with the corresponding normal operation data interval of the wind turbine equipment, and whether the wind turbine equipment fails is determined according to the matching result. When the wind turbine equipment fails, the specific steps of generating abnormal operation data of the wind turbine equipment are as follows:

[0065] S51. Numerically match each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set. If each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set is successfully matched with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set, it is determined that the wind turbine equipment has not failed and no operation is performed; otherwise, it is determined that the wind turbine equipment has failed, and the wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set that fails to match the corresponding wind turbine equipment normal operation data interval is screened out to generate a wind turbine equipment abnormal operation data set. ,in, Indicates abnormal operation data of wind turbine equipment, Indicates the total number of abnormal operation data of wind turbine equipment.

[0066] The standard operation data of the wind turbine equipment is numerically matched with the set normal operation data interval of the wind turbine equipment, and whether the wind turbine equipment fails is determined based on the matching result. When the wind turbine fails, abnormal operation data of the wind turbine equipment is generated, thereby ensuring the accuracy of the determination result and realizing intelligent detection of wind turbine faults.

[0067] Preferably, the specific steps of generating corresponding abnormal wind turbine device text data according to the abnormal operation data of the wind turbine device are as follows:

[0068] S61, using a bidirectional search algorithm to search for wind turbine equipment that matches the abnormal operation data of each wind turbine equipment in the wind turbine equipment abnormal operation data set from the wind turbine equipment set, and generate a wind turbine equipment abnormal equipment text data set through data identification ,in, Indicates The abnormal equipment text data of wind turbines corresponding to the abnormal operation data of wind turbines, Indicates the total number of abnormal equipment text data of wind turbines.

[0069] A bidirectional search algorithm is used to search for wind turbine equipment corresponding to abnormal operation data of wind turbine equipment and generate abnormal wind turbine equipment text data through data identification, thereby improving the speed and efficiency of the search process.

[0070] Preferably, the specific steps of constructing wind turbine equipment failure prediction result data and performing wind turbine equipment failure early warning operations are as follows:

[0071] S71, the abnormal equipment text data set of the wind turbine Dataset of abnormal operation of wind turbine equipment Combine data to construct wind turbine equipment failure prediction result data ;

[0072] S72. Push the wind turbine equipment fault prediction result data to the wind turbine equipment monitoring unit via the wireless communication network, and issue an early warning using a signal light.

[0073] Construct wind turbine equipment failure prediction result data and execute wind turbine equipment failure warning operations to achieve real-time warning of wind turbine equipment.

[0074] The present invention also discloses a data-driven wind turbine equipment fault prediction system, comprising a real-time operation data acquisition module, a real-time operation data cleaning module, a real-time operation data numerical processing module, a normal operation data interval setting module, an abnormal operation data determination module and a text data generation module;

[0075] The real-time operation data acquisition module collects the real-time operation data of the wind turbine equipment through the equipment monitoring device according to the set wind turbine equipment set and the operation data acquisition time point set;

[0076] The real-time operation data cleaning module uses the DBSCAN algorithm to clean the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment;

[0077] The real-time operation data numerical processing module normalizes the real operation data of the wind turbine equipment by the maximum and minimum normalization method to obtain the normalized operation data of the wind turbine equipment, and then uses the mean formula to perform average numerical measurement processing on the normalized operation data of the wind turbine equipment to generate standard operation data of the wind turbine equipment;

[0078] The normal operating data interval setting module sets the normal operating data interval set of the wind turbine equipment by using an artificial hummingbird optimization algorithm;

[0079] The abnormal operation data determination module numerically matches the standard operation data of each wind turbine equipment with the normal operation data interval of the wind turbine equipment corresponding to the normal operation data interval set of the wind turbine equipment, determines whether the wind turbine equipment fails according to the matching result, and generates abnormal operation data of the wind turbine equipment when the wind turbine equipment fails;

[0080] The text data generation module uses a bidirectional search algorithm to search for wind turbine equipment that matches abnormal operation data of wind turbine equipment from the wind turbine equipment set and generates abnormal wind turbine equipment text data through data identification.

[0081] (III) Beneficial effects

[0082] 1. The present invention clusters the collected real-time operation data through the DBSCAN algorithm, removes the outlier data therein, obtains the real operation data of the wind turbine equipment, and effectively avoids the influence of the outlier data; normalizes the real operation data of the wind turbine equipment through the maximum and minimum standardization method, and then uses the mean formula to average the normalized data to obtain the standard operation data of the wind turbine equipment, thereby ensuring the accuracy of the acquired data and reducing the calculation amount of subsequent operations; numerically matches the standard operation data of the wind turbine equipment with the set normal operation data interval of the wind turbine equipment, and determines whether the wind turbine equipment fails according to the matching result. When the wind turbine equipment fails, abnormal operation data of the wind turbine equipment is generated, thereby ensuring the accuracy of the determination result and realizing intelligent detection of wind turbine faults; a two-way search algorithm is used to search for the wind turbine equipment corresponding to the abnormal operation data of the wind turbine equipment and generate abnormal equipment text data of the wind turbine equipment through data identification, thereby improving the speed and efficiency of the search process; constructs wind turbine equipment fault prediction result data and executes wind turbine equipment fault warning operations to realize real-time warning of wind turbine equipment;

[0083] 2. Compared with analyzing and processing the collected historical data to generate the normal operating data interval of the wind turbine equipment, the present invention sets the normal operating data interval of the wind turbine equipment through the artificial hummingbird optimization algorithm, which improves the accuracy of the setting results, speeds up the optimization process, and effectively avoids the optimization results from falling into the local optimum. BRIEF DESCRIPTION OF THE DRAWINGS

[0084] In order to more clearly illustrate the technical solutions of the embodiments of the invention, the drawings required for describing the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the invention. For ordinary technicians in this field, they can also obtain drawings based on these drawings without paying creative work.

[0085] Figure 1 A flow chart of a data-driven wind turbine equipment fault prediction method provided by the present invention;

[0086] Figure 2 A module schematic diagram of a data-driven wind turbine equipment fault prediction system provided by the present invention. DETAILED DESCRIPTION

[0087] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0088] In the description of the present invention, it is necessary to understand that the terms "opening", "upper", "lower", "top", "middle", "inside" and the like indicating orientation or positional relationship are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the components or elements referred to must have a specific orientation, be constructed and operate in a specific orientation, and therefore cannot be understood as limiting the invention.

[0089] Embodiment 1 is as follows:

[0090] See also Figure 1 , a data-driven wind turbine equipment fault prediction method, comprising the following steps:

[0091] S1. Collect real-time operation data of wind turbine equipment through equipment monitoring devices;

[0092] S11. Set wind turbine equipment set ,in, Indicates Wind turbine operating equipment, Indicates the total number of wind turbine operating equipment;

[0093] S12: Set an operation data collection cycle, divide the operation data collection cycle evenly, and obtain an operation data collection time point set. ,in, Indicates The operation data collection time point, Indicates the total number of time points for running data collection;

[0094] S13, collecting the wind turbine equipment real-time operation data of each wind turbine equipment at each operation data collection time point through the equipment monitoring device, and obtaining the wind turbine equipment real-time operation data matrix as follows:

[0095] ,

[0096] in, Indicates that the equipment monitoring device is The first operation data collection time point is collected Real-time operating data of wind turbine operating equipment.

[0097] S2. Using a clustering algorithm to clean up the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment;

[0098] S21, using DBSCAN algorithm to analyze the real-time operation data matrix of wind turbine equipment The real-time operation data of each wind turbine operating equipment are clustered separately, and the real-time operation data matrix of wind turbine equipment is eliminated according to the clustering results. The outlier data in the real-time operation data of each wind turbine operating equipment is obtained to obtain the real operation data matrix of the wind turbine equipment. as follows:

[0099] ,

[0100] in, Indicates The first wind turbine operating equipment The real operating data of wind turbine equipment, It represents the total number of real operation data of wind turbine equipment in the real-time operation data of wind turbine equipment collected by the equipment monitoring device; the clustering process is as follows:

[0101] S211, set the neighborhood radius to And the data volume threshold in the core point neighborhood is ;

[0102] S212, randomly select a wind turbine operating device, and generate a real-time operating data matrix of the wind turbine device All real-time operating data of the wind turbine operating equipment described in the above are taken as a data set;

[0103] S213: Select any data in the data set, calculate the Euclidean distance between the data set and other data in the data set, and calculate the Euclidean distance less than the neighborhood radius. The total number of ;

[0104] S214, Comparison and If the value of , then the selected data is determined to be a core point; otherwise, the selected data is determined to be an outlier;

[0105] S215, repeating the operations from S213 to S214 until all the data in the data set are traversed, and then eliminating outlier points in the data set that are not within the neighborhood radius of any core point.

[0106] S3, performing numerical processing on the actual operation data of the wind turbine generator set equipment to obtain standard operation data of the wind turbine generator set equipment;

[0107] S31, the actual operation data matrix of the wind turbine equipment is normalized by the maximum and minimum standardization method. The real operation data of the wind turbine equipment in the normalization process is processed to obtain the normalized operation data matrix of the wind turbine equipment as follows:

[0108] ,

[0109] in, Indicates the actual operation data matrix of wind turbine equipment The first wind turbine operating equipment The normalized operating data of wind turbine equipment is obtained after normalization of the actual operating data of each wind turbine equipment; the calculation formula is as follows:

[0110] ,

[0111] in, and They represent the real operation data matrix of wind turbine equipment. The maximum and minimum values ​​of the actual operation data of all wind turbine equipment in the wind turbine operation equipment;

[0112] S32, using the mean formula to normalize the operation data matrix of the wind turbine equipment The normalized operation data of the wind turbine equipment in the system is processed by average numerical measurement to obtain the standard operation data set of the wind turbine equipment. ,in, Indicates The wind turbine equipment standard operating data of each wind turbine operating equipment, and , Indicates the total number of wind turbine operating devices.

[0113] S4. Setting a normal operating data interval set of wind turbine equipment through an intelligent optimization algorithm;

[0114] S41. Build a hummingbird population and set the hummingbird population size to , the current number of iterations is The maximum number of iterations is And the dimension of the running data interval search space is ;

[0115] Randomly select a wind turbine device and generate a random value between [0,1] operating data intervals as normal operating data interval search samples for the wind turbine device, each operating data interval is used as a food source, and each hummingbird in the hummingbird population is assigned to a food source;

[0116] Randomly generate the initial position set of the hummingbird population in the running data interval search space ,in, Indicates the first The initial positions of individual hummingbirds;

[0117] S42, calculating the fitness function value of the food source where each hummingbird individual in the hummingbird population is located; the fitness function formula is as follows:

[0118] ,

[0119] in, Indicates the first The fitness function value of the food source where the hummingbird individual is located, Indicates the first The operating data interval corresponding to the food source where the hummingbird individuals are located meets the normal working capacity of the wind turbine equipment. Indicates the correction value;

[0120] S43. Each hummingbird in the hummingbird population will randomly select any one of the flight modes of axial flight, omnidirectional flight and diagonal flight to conduct guided foraging or territorial foraging in the search space of the running data interval;

[0121] S431. When a hummingbird individual is guided to forage in the search space of the running data interval, the hummingbird individual will select the food source where other hummingbird individuals are located as the target food source according to probability; the probability of each hummingbird individual being selected is as follows:

[0122] ,

[0123] in, Indicates The probability that the food source where the hummingbird individual is located is selected as the target food source;

[0124] After a hummingbird has selected a target food source, it will update its position according to the location of the target food source. The position update formula is as follows:

[0125] ,

[0126] in, Indicates the first The hummingbird individuals The location of the candidate food source selected after the location update during the iteration; Indicates the first The hummingbird individuals The position during the iteration; Indicates the first The hummingbird individuals The location of the target food source selected during the iteration; represents the guiding factor, which follows the normal distribution of (0,1); represents the flight mode chosen by individual hummingbirds;

[0127] S432. When a hummingbird individual is foraging in a territory in the search space of the running data interval, the hummingbird individual will search for a new food source near the food source to update its position; the position update formula is as follows:

[0128] ,

[0129] in, represents the regional factor, which follows the normal distribution of (0,1);

[0130] S44, calculating the fitness function value of the candidate food source selected by each hummingbird individual in the hummingbird population, if the fitness function value of the candidate food source selected by the hummingbird individual is greater than the fitness function value of the original food source, then the candidate food source is used to replace the original food source; otherwise, the original food source is retained;

[0131] S45, setting a migration threshold. If the current number of iterations is greater than or equal to the migration threshold, the hummingbird individual at the food source with the smallest fitness function value will continue to migrate for food, abandon the current food source, and randomly search for a new food source in the running data interval search space, and calculate the fitness function value of the food source.

[0132] S46, arranging the food sources of all hummingbird individuals according to the fitness function values, and selecting the food source with the highest fitness function value as the current optimal solution;

[0133] S47, determine the current number of iterations Is it less than the maximum number of iterations? , if the current number of iterations Less than the maximum number of iterations , then the current number of iterations Add 1 and return to S43; otherwise, take the current optimal solution as the global optimal solution, and output the operating data interval corresponding to the global optimal solution as the normal operating data interval of the selected wind turbine equipment;

[0134] S48, repeat the operations from S41 to S47 to generate a normal operating data interval set of the wind turbine equipment. ,in, Indicates The normal operating data range of each wind turbine equipment.

[0135] S5, numerically matching the standard operation data of the wind turbine generator set equipment with the corresponding normal operation data interval of the wind turbine generator set equipment, determining whether the wind turbine generator set equipment fails according to the matching result, and generating abnormal operation data of the wind turbine generator set equipment when the wind turbine generator set equipment fails;

[0136] S51. Numerically match each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set. If each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set is successfully matched with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set, it is determined that the wind turbine equipment has not failed and no operation is performed; otherwise, it is determined that the wind turbine equipment has failed, and the wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set that fails to match the corresponding wind turbine equipment normal operation data interval is screened out to generate a wind turbine equipment abnormal operation data set. ,in, Indicates abnormal operation data of wind turbine equipment, Indicates the total number of abnormal operation data of wind turbine equipment.

[0137] S6. Generate corresponding wind turbine equipment text data according to the abnormal operation data of the wind turbine equipment;

[0138] S61, using a bidirectional search algorithm to search for wind turbine equipment that matches the abnormal operation data of each wind turbine equipment in the wind turbine equipment abnormal operation data set from the wind turbine equipment set, and generate a wind turbine equipment abnormal equipment text data set through data identification ,in, Indicates The abnormal equipment text data of wind turbines corresponding to the abnormal operation data of wind turbines, Indicates the total number of abnormal equipment text data of wind turbines.

[0139] S7, construct wind turbine equipment failure prediction result data, and perform wind turbine equipment failure warning operations;

[0140] S71, the abnormal equipment text data set of the wind turbine Dataset of abnormal operation of wind turbine equipment Combine data to construct wind turbine equipment failure prediction result data ;

[0141] S72. Push the wind turbine equipment fault prediction result data to the wind turbine equipment monitoring unit via the wireless communication network, and issue an early warning using a signal light.

[0142] Embodiment 2 is as follows:

[0143] See also Figure 2 , a system based on a data-driven wind turbine equipment fault prediction method, comprising a real-time operation data acquisition module, a real-time operation data cleaning module, a real-time operation data numerical processing module, a normal operation data interval setting module, an abnormal operation data determination module and a text data generation module;

[0144] The real-time operation data acquisition module collects the real-time operation data of the wind turbine equipment through the equipment monitoring device according to the set wind turbine equipment set and the operation data acquisition time point set;

[0145] The real-time operation data cleaning module uses the DBSCAN algorithm to clean the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment;

[0146] The real-time operation data numerical processing module normalizes the real operation data of the wind turbine equipment by the maximum and minimum normalization method to obtain the normalized operation data of the wind turbine equipment, and then uses the mean formula to perform average numerical measurement processing on the normalized operation data of the wind turbine equipment to generate standard operation data of the wind turbine equipment;

[0147] The normal operating data interval setting module sets the normal operating data interval set of the wind turbine equipment by using an artificial hummingbird optimization algorithm;

[0148] The abnormal operation data determination module numerically matches the standard operation data of each wind turbine equipment with the normal operation data interval of the wind turbine equipment corresponding to the normal operation data interval set of the wind turbine equipment, determines whether the wind turbine equipment fails according to the matching result, and generates abnormal operation data of the wind turbine equipment when the wind turbine equipment fails;

[0149] The text data generation module uses a bidirectional search algorithm to search for wind turbine equipment that matches abnormal operation data of wind turbine equipment from the wind turbine equipment set and generates abnormal wind turbine equipment text data through data identification.

[0150] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0151] The preferred embodiments of the invention disclosed above are only used to help explain the invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the invention, so that those skilled in the art can understand and use the invention well.

Claims

1. A data-driven wind turbine equipment fault prediction method, characterized in that: The steps include: S1. Collect real-time operation data of wind turbine equipment through equipment monitoring devices; S2. Using a clustering algorithm to clean up the real-time operation data of the wind turbine equipment to obtain the real operation data of the wind turbine equipment; S3, performing numerical processing on the actual operation data of the wind turbine equipment to obtain standard operation data of the wind turbine equipment; S4. Setting a normal operating data interval set of wind turbine equipment through an intelligent optimization algorithm; S5, numerically matching the standard operation data of the wind turbine generator set equipment with the corresponding normal operation data interval of the wind turbine generator set equipment, determining whether the wind turbine generator set equipment fails according to the matching result, and generating abnormal operation data of the wind turbine generator set equipment when the wind turbine generator set equipment fails; S6. Generate corresponding wind turbine equipment text data according to the abnormal operation data of the wind turbine equipment; S7, construct wind turbine equipment failure prediction result data, and perform wind turbine equipment failure warning operations; The S4 comprises the following steps: S41. Build a hummingbird population, set the hummingbird population size to N, the current number of iterations to t, and the maximum number of iterations to t max And the dimension of the running data interval search space is M; A wind turbine device is randomly selected, and N operating data intervals are randomly generated between [0, 1] as normal operating data interval search samples of the wind turbine device, each operating data interval is used as a food source, and each hummingbird in the hummingbird population is assigned to a food source; An initial position set of the hummingbird population is randomly generated in the running data interval search space; S42, calculating the fitness function value of the food source where each hummingbird individual in the hummingbird population is located; S43. Each hummingbird in the hummingbird population will randomly select any one of the flight modes of axial flight, omnidirectional flight and diagonal flight to conduct guided foraging or territorial foraging in the search space of the running data interval; S431, when the hummingbird individual is guided to forage in the search space of the running data interval, the hummingbird individual will select the food source where other hummingbird individuals are located as the target food source according to probability; When a hummingbird individual selects a target food source, it will update its position according to the location of the target food source; S432, when the hummingbird individual is foraging in the territory in the search space of the running data interval, the hummingbird individual will search for a new food source near the food source to update the position; S44, calculating the fitness function value of the candidate food source selected by each hummingbird individual in the hummingbird population, if the fitness function value of the candidate food source selected by the hummingbird individual is greater than the fitness function value of the original food source, then the candidate food source is used to replace the original food source; otherwise, the original food source is retained; S45, setting a migration threshold. If the current number of iterations is greater than or equal to the migration threshold, the hummingbird individual at the food source with the smallest fitness function value will continue to migrate for food, abandon the current food source, and randomly search for a new food source in the running data interval search space, and calculate the fitness function value of the food source. S46, arranging the food sources of all hummingbird individuals according to the fitness function values, and selecting the food source with the highest fitness function value as the current optimal solution; S47, determine whether the current number of iterations t is less than the maximum number of iterations t max , if the current number of iterations t is less than the maximum number of iterations t max , then the current iteration number t is increased by 1, and the process returns to S43; otherwise, the current optimal solution is taken as the global optimal solution, and the operating data interval corresponding to the global optimal solution is output as the normal operating data interval of the selected wind turbine equipment; S48. Repeat the operations from S41 to S47 to generate a normal operation data interval set of the wind turbine equipment.

2. The data-driven wind turbine equipment fault prediction method according to claim 1, characterized in that: The S1 comprises the following steps: S11, set the wind turbine equipment set A = {a1, a2, ..., a i ,…,a k }, where a i represents the operating equipment of the i-th wind turbine, and k represents the total number of operating equipment of the wind turbine; S12, setting an operation data collection cycle, and evenly dividing the operation data collection cycle to obtain an operation data collection time point set B = {b1, b2, ..., b i ,…,b l }, where b i represents the i-th operation data collection time point, l represents the total number of operation data collection time points; S13, collecting the wind turbine equipment real-time operation data of each wind turbine equipment at each operation data collection time point through the equipment monitoring device, and obtaining the wind turbine equipment real-time operation data matrix C.

3. The data-driven wind turbine equipment fault prediction method according to claim 2, characterized in that: The S2 comprises the following steps: S21. Use the DBSCAN algorithm to cluster the real-time operation data of each wind turbine operating device in the real-time operation data matrix C of wind turbine equipment respectively, and remove the outlier data in the real-time operation data of each wind turbine operating device in the real-time operation data matrix C of wind turbine equipment according to the clustering results to obtain the real operation data matrix C' of wind turbine equipment.

4. A data-driven wind turbine equipment fault prediction method according to claim 3, characterized in that: The clustering process in S21 includes the following steps: S211, setting the neighborhood radius to r and the data volume threshold in the core point neighborhood to h; S212, randomly selecting a wind turbine operating device, and taking all real-time operating data of the wind turbine operating device in the wind turbine device real-time operating data matrix C as a data set; S213, select any data in the data set, calculate the Euclidean distance between the data set and other data in the data set, and count the total number h1 of data whose calculated Euclidean distance is less than the neighborhood radius r; S214, compare the numerical values ​​of h1 and h, if h1 ≥ h, determine that the selected data is a core point; otherwise, determine that the selected data is an outlier; S215, repeating the operations from S213 to S214 until all the data in the data set are traversed, and then eliminating outlier points in the data set that are not within the neighborhood radius of any core point.

5. A data-driven wind turbine equipment fault prediction method according to claim 4, characterized in that: The S3 comprises the following steps: S31, performing operation data normalization processing on the wind turbine equipment real operation data in the wind turbine equipment real operation data matrix C' by using the maximum and minimum normalization method to obtain the wind turbine equipment normalized operation data matrix S32, using the mean formula to normalize the operation data matrix of the wind turbine equipment The normalized operation data of the wind turbine equipment in the system are processed by average numerical measurement to obtain the standard operation data set of the wind turbine equipment.

6. A data-driven wind turbine equipment fault prediction method according to claim 5, characterized in that: The S5 comprises the following steps: S51. Numerically match each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set. If each wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set is successfully matched with the wind turbine equipment normal operation data interval corresponding to the wind turbine equipment normal operation data interval set, it is determined that the wind turbine equipment has not failed and no operation is performed; otherwise, it is determined that the wind turbine equipment has failed, and the wind turbine equipment standard operation data set in the wind turbine equipment standard operation data set that fails to match the corresponding wind turbine equipment normal operation data interval is screened out to generate a wind turbine equipment abnormal operation data set. in, represents the abnormal operation data of the i-th wind turbine equipment, and p represents the total number of abnormal operation data of the wind turbine equipment.

7. A data-driven wind turbine equipment fault prediction method according to claim 6, characterized in that: The S6 comprises the following steps: S61, using a bidirectional search algorithm to search for wind turbine equipment that matches the abnormal operation data of each wind turbine equipment in the wind turbine equipment abnormal operation data set from the wind turbine equipment set, and generate a wind turbine equipment abnormal equipment text data set A'={a'1, a'2, ..., a' i ,…,a' p }, where a' i represents the abnormal equipment text data of the wind turbine set corresponding to the abnormal operation data of the i-th wind turbine set equipment, and p represents the total number of abnormal equipment text data of the wind turbine set.

8. The data-driven wind turbine equipment fault prediction method according to claim 7, characterized in that: The S7 comprises the following steps: S71, combining the abnormal equipment text data set A' of the wind turbine generator set with the abnormal operation data set of the wind turbine generator set Combine data to construct wind turbine equipment failure prediction result data S72. Push the wind turbine equipment fault prediction result data to the wind turbine equipment monitoring unit via the wireless communication network, and issue an early warning using a signal light.

9. A system for implementing the data-driven wind turbine equipment fault prediction method as described in any one of claims 1 to 8, comprising a real-time operation data acquisition module, a real-time operation data cleaning module, a real-time operation data numerical processing module, a normal operation data interval setting module, an abnormal operation data determination module and a text data generation module.

Images