A bearing wear detection method and device for wind turbines

Through the unsupervised model, abnormal wind turbines were screened out, and the problems of low detection efficiency and poor accuracy in the existing technology were solved, and efficient and accurate bearing detection was achieved.

CN113902943BActive Publication Date: 2024-12-06新疆华电苇湖梁新能源有限公司

Patent Information

Application Number
CN202111155772.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-09-29
Publication Date
2024-12-06
Estimated Expiration
2041-09-29

AI Technical Summary

Technical Problem

The prior art is difficult to accurately detect the wear of wind turbine bearings, resulting in low detection efficiency and poor accuracy, which requires manual observation or collection of grease for analysis, which is time-consuming and labor-consuming.

Method used

Unsupervised models using multiple abnormal data detection methods are used to analyze the bearing temperature data of wind turbines, and abnormal wind turbines are screened through abnormal probability thresholds to improve the accuracy and efficiency of detection.

Benefits of technology

Accurate detection of the bearing status of the wind turbine is achieved without manual entry of the generator, reducing time and energy consumption, and improving detection efficiency and accuracy.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN113902943B_ABST
    Figure CN113902943B_ABST
Patent Text Reader

Abstract

The embodiment of the present application discloses a bearing detection method and related device for a wind turbine, which can first obtain the bearing temperature data corresponding to the target wind turbine group, and the bearing temperature data is used to reflect the bearing temperature of each wind turbine in the target wind turbine group during operation. Through multiple unsupervised models corresponding to multiple abnormal data detection methods, the bearing temperature data can be analyzed from multiple angles, so as to more accurately obtain the normal state of the bearing temperature of the target wind turbine group during operation. The abnormal bearing temperature data can reflect the probability of abnormality of the wind turbine bearing, and then can realize accurate detection of wind turbines with abnormal bearings, without the need for manual entry into the wind turbine, thereby improving the efficiency and accuracy of bearing detection and reducing the consumption of time and energy.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field, and in particular to a bearing wear detection method and related device for a wind turbine. Background Art

[0002] Wind power generation is one of the main ways to obtain clean energy at present. Wind energy can be converted into electrical energy for use through wind turbines. Among them, the wind turbine mainly connects the impeller and the generator through bearings. When the wind blows the blades of the wind turbine, the impeller of the blade can drive the rotor of the generator to rotate through the bearings to generate electricity. It can be seen that the bearing status in the wind turbine is closely related to the operating status of the wind turbine.

[0003] In related technologies, in order to ensure the stable operation of wind turbines, relevant personnel are required to enter the wind turbine and observe the bearing conditions with the naked eye, such as the number and size of cracks on the bearings, or collect grease in the wind turbine for analysis, such as observing the grease content and the iron filings content in the grease.

[0004] This detection method requires a lot of energy and time, and cannot accurately analyze the bearing condition, and the detection effect on the bearing is poor. Summary of the invention

[0005] In order to solve the above technical problems, the present application provides a bearing wear detection method for wind turbines. The processing equipment can analyze the collected bearing temperature data of the target wind turbine generator set according to multiple unsupervised models with multiple abnormal data detection methods, so as to more accurately determine the wind turbines with abnormal bearing temperature data, thereby improving the accuracy of detecting the bearing status of the wind turbine.

[0006] The embodiments of the present application disclose the following technical solutions:

[0007] In a first aspect, an embodiment of the present application discloses a bearing detection method for a wind turbine, the method comprising:

[0008] Acquire bearing temperature data corresponding to the target wind turbine generator set, wherein the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation;

[0009] Determining, by using a plurality of unsupervised models, a first abnormal probability of each generator in the target wind turbine generator set in a bearing temperature dimension according to the bearing temperature data, wherein the plurality of unsupervised models correspond to a plurality of abnormal data detection methods;

[0010] A wind turbine generator having a first abnormal probability greater than a first probability threshold is determined as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0011] In a possible implementation, the step of obtaining bearing temperature data corresponding to the target wind turbine generator set includes:

[0012] Acquiring operation data corresponding to the target wind turbine generator set within a preset period of time, wherein the operation data includes pending bearing temperature data corresponding to the target wind turbine generator set within the preset period of time, and associated data associated with the pending bearing temperature data, wherein the associated data includes any one or more combinations of wind turbine status data, active power data, wind turbine speed data, cabin temperature data, and cabin vibration data;

[0013] In response to the data volume of the operating data being greater than a data volume threshold and the number of wind turbines in the target wind turbine generator set being greater than a turbine number threshold, filtering the operating data based on the associated data and a data filtering threshold, wherein the data filtering threshold is used to reflect the normal operating status of the wind turbines in the target wind turbine generator set;

[0014] The undetermined bearing temperature data included in the screened operating data is determined as the bearing temperature data corresponding to the target wind turbine generator set.

[0015] In a possible implementation manner, the operation data includes the cabin vibration data, and the method further includes:

[0016] Determining the nacelle vibration data associated with the bearing temperature data as the target nacelle vibration data corresponding to the target wind turbine generator set;

[0017] Determining, by means of the multiple unsupervised models and based on the target nacelle vibration data, a second abnormal probability of each generator in the target wind turbine generator set in a bearing vibration dimension;

[0018] The step of determining the wind turbine generator having a first abnormal probability greater than a first probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group includes:

[0019] A wind turbine generator having a first abnormal probability greater than a first probability threshold and a second abnormal probability greater than a second probability threshold is determined as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0020] In a possible implementation, the operation data includes the fan speed data, and the method further includes:

[0021] Determining the wind turbine speed data associated with the bearing temperature data as the target wind turbine speed data corresponding to the target wind turbine generator set;

[0022] grouping the bearing temperature data and the target nacelle vibration data according to the target fan speed data;

[0023] A wind turbine generator corresponding to at least one of the first abnormal probability and the second abnormal probability in at least one speed group is determined to be a wind turbine generator with a bearing abnormality in the target wind turbine generator group.

[0024] In a possible implementation, determining a wind turbine generator having a first abnormal probability greater than a first probability threshold and a second abnormal probability greater than a second probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group includes:

[0025] A wind turbine having a first abnormality probability obtained in at least one unsupervised model greater than a first probability threshold and a second abnormality probability obtained in at least one unsupervised model greater than a second probability threshold is determined as a wind turbine having a bearing abnormality in the target wind turbine group.

[0026] In a possible implementation, the wind turbine in the target wind turbine generator set is a direct-drive wind turbine generator, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data, and second bearing generator side temperature data, and the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the bearing temperature data by using multiple unsupervised models, including:

[0027] Determine first temperature data according to an average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0028] determining second temperature data according to a difference between a maximum value among the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data;

[0029] Determine third temperature data according to an average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data;

[0030] Determine fourth temperature data according to an average value of the first bearing generator side temperature data and the second bearing generator side temperature data;

[0031] determining fifth temperature data according to a difference between the third temperature data and the fourth temperature data;

[0032] Determine sixth temperature data according to a standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0033] Determine a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data;

[0034] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the first temperature data matrix through the multiple unsupervised models.

[0035] In a possible implementation, the wind turbine in the target wind turbine generator set is a non-direct drive wind turbine generator, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes gearbox high-speed shaft front shaft temperature data and gearbox high-speed shaft rear shaft temperature data, and the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the bearing temperature data by using multiple unsupervised models, including:

[0036] Determine seventh temperature data according to the difference between the temperature data of the front axle of the gearbox high-speed shaft and the temperature data of the rear axle of the gearbox high-speed shaft;

[0037] Determine eighth temperature data according to the difference between the gearbox high-speed shaft rear axle temperature data and the cabin temperature data;

[0038] Determine the gearbox high-speed shaft rear axle temperature data as ninth temperature data;

[0039] Determine a second temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the seventh temperature data, the eighth temperature data and the ninth temperature data;

[0040] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the second temperature data matrix through the multiple unsupervised models.

[0041] In a possible implementation, the target nacelle vibration data includes nacelle transverse axis vibration data and nacelle longitudinal axis vibration data, and determining the second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension according to the target nacelle vibration data by using the multiple unsupervised models includes:

[0042] determining first vibration data according to a difference between an absolute value of the cabin transverse axis vibration data and an absolute value of a mean value of the cabin transverse axis vibration data;

[0043] determining second vibration data according to a difference between an absolute value of the cabin longitudinal axis vibration data and an absolute value of a mean value of the cabin longitudinal axis vibration data;

[0044] Determine a vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data;

[0045] The second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension is determined according to the vibration data matrix through the multiple unsupervised models.

[0046] In a possible implementation, the wind turbine in the target wind turbine generator set is a non-direct drive wind turbine, and the method further includes:

[0047] determining third vibration data according to a maximum value of the cabin transverse axis vibration data and the cabin longitudinal axis vibration data;

[0048] Determining the vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data includes:

[0049] A vibration data matrix corresponding to each wind turbine in the target wind turbine generator group is determined according to the first vibration data, the second vibration data and the third vibration data.

[0050] In a second aspect, an embodiment of the present application discloses a bearing detection device for a wind turbine generator, the device comprising an acquisition unit, a first determination unit, and a second determination unit:

[0051] An acquisition unit, used for acquiring bearing temperature data corresponding to a target wind turbine generator set, wherein the bearing temperature data is used for reflecting the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation;

[0052] A first determination unit is configured to determine, based on the bearing temperature data, a first abnormal probability of each generator in the target wind turbine generator set in a bearing temperature dimension by using a plurality of unsupervised models, wherein the plurality of unsupervised models correspond to a plurality of abnormal data detection methods;

[0053] The second determining unit is configured to determine a wind turbine generator having a first abnormality probability greater than a first probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0054] In a possible implementation, the acquiring unit is specifically configured to:

[0055] Acquiring operation data corresponding to the target wind turbine generator set within a preset period of time, wherein the operation data includes pending bearing temperature data corresponding to the target wind turbine generator set within the preset period of time, and associated data associated with the pending bearing temperature data, wherein the associated data includes any one or more combinations of wind turbine status data, active power data, wind turbine speed data, cabin temperature data, and cabin vibration data;

[0056] In response to the data volume of the operating data being greater than a data volume threshold and the number of wind turbines in the target wind turbine generator set being greater than a turbine number threshold, filtering the operating data based on the associated data and a data filtering threshold, wherein the data filtering threshold is used to reflect the normal operating status of the wind turbines in the target wind turbine generator set;

[0057] The undetermined bearing temperature data included in the screened operating data is determined as the bearing temperature data corresponding to the target wind turbine generator set.

[0058] In a possible implementation manner, the operation data includes the cabin vibration data, and the device further includes a third determining unit and a fourth determining unit:

[0059] A third determining unit, configured to determine the nacelle vibration data associated with the bearing temperature data as target nacelle vibration data corresponding to the target wind turbine generator set;

[0060] a fourth determination unit, configured to determine, by using the multiple unsupervised models and according to the target nacelle vibration data, a second abnormal probability of each generator in the target wind turbine generator set in a bearing vibration dimension;

[0061] The second determining unit is specifically configured to:

[0062] A wind turbine generator having a first abnormal probability greater than a first probability threshold and a second abnormal probability greater than a second probability threshold is determined as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0063] In a possible implementation, the operation data includes the fan speed data, and the device further includes a fifth determination unit, a grouping unit, and a sixth determination unit:

[0064] a fifth determining unit, configured to determine the wind turbine speed data associated with the bearing temperature data as the target wind turbine speed data corresponding to the target wind turbine generator set;

[0065] A grouping unit, configured to group the bearing temperature data and the target nacelle vibration data according to the target fan speed data;

[0066] The sixth determination unit is used to determine, in at least one speed group, a wind turbine corresponding to at least one of the first abnormal probability and the second abnormal probability greater than a third probability threshold as a wind turbine with a bearing abnormality in the target wind turbine group.

[0067] In a possible implementation manner, the second determining unit is specifically configured to:

[0068] A wind turbine having a first abnormality probability obtained in at least one unsupervised model greater than a first probability threshold and a second abnormality probability obtained in at least one unsupervised model greater than a second probability threshold is determined as a wind turbine having a bearing abnormality in the target wind turbine group.

[0069] In a possible implementation, the wind turbine in the target wind turbine generator set is a direct-drive wind turbine, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data, and second bearing generator side temperature data, and the first determining unit 702 is specifically used to:

[0070] Determine first temperature data according to an average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0071] determining second temperature data according to a difference between a maximum value among the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data;

[0072] Determine third temperature data according to an average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data;

[0073] Determine fourth temperature data according to an average value of the first bearing generator side temperature data and the second bearing generator side temperature data;

[0074] determining fifth temperature data according to a difference between the third temperature data and the fourth temperature data;

[0075] Determine sixth temperature data according to a standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0076] Determine a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data;

[0077] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the first temperature data matrix through the multiple unsupervised models.

[0078] In a possible implementation, the wind turbine in the target wind turbine generator set is a non-direct drive wind turbine, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes gearbox high-speed shaft front shaft temperature data and gearbox high-speed shaft rear shaft temperature data, and the first determination unit is specifically used to:

[0079] Determine seventh temperature data according to the difference between the temperature data of the front axle of the gearbox high-speed shaft and the temperature data of the rear axle of the gearbox high-speed shaft;

[0080] Determine eighth temperature data according to the difference between the gearbox high-speed shaft rear axle temperature data and the cabin temperature data;

[0081] Determine the gearbox high-speed shaft rear axle temperature data as ninth temperature data;

[0082] Determine a second temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the seventh temperature data, the eighth temperature data and the ninth temperature data;

[0083] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the second temperature data matrix through the multiple unsupervised models.

[0084] In a possible implementation, the target cabin vibration data includes cabin transverse axis vibration data and cabin longitudinal axis vibration data, and the fourth determination unit is specifically configured to:

[0085] determining first vibration data according to a difference between an absolute value of the cabin transverse axis vibration data and an absolute value of a mean value of the cabin transverse axis vibration data;

[0086] determining second vibration data according to a difference between an absolute value of the cabin longitudinal axis vibration data and an absolute value of a mean value of the cabin longitudinal axis vibration data;

[0087] Determine a vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data;

[0088] The second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension is determined according to the vibration data matrix through the multiple unsupervised models.

[0089] In a possible implementation, the wind turbine in the target wind turbine generator group is a non-direct drive wind turbine, and the method further includes a seventh determining unit:

[0090] a seventh determination unit, configured to determine third vibration data according to a maximum value of the cabin transverse axis vibration data and the cabin longitudinal axis vibration data;

[0091] The fourth determining unit is specifically configured to:

[0092] A vibration data matrix corresponding to each wind turbine in the target wind turbine generator group is determined according to the first vibration data, the second vibration data and the third vibration data.

[0093] It can be seen from the above technical solution that when detecting the bearing, the bearing temperature data corresponding to the target wind turbine generator set can be obtained first, and the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation. Through multiple unsupervised models corresponding to multiple abnormal data detection methods, the bearing temperature data can be analyzed from multiple angles, so as to more accurately obtain the normal state of the bearing temperature of the target wind turbine generator set during operation, and then the more abnormal temperature data can be determined from the bearing temperature data. The abnormal bearing temperature data can reflect the probability of abnormality of the wind turbine generator bearing. By determining the wind turbine generator with a first abnormal probability greater than the first probability threshold as a bearing abnormal wind turbine generator in the target wind turbine generator set, accurate detection of wind turbine generators with abnormal bearings can be achieved without manual entry into the wind turbine generator, thereby improving the efficiency and accuracy of bearing detection and reducing the consumption of time and energy. BRIEF DESCRIPTION OF THE DRAWINGS

[0094] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0095] Figure 1 A flow chart of a bearing detection method for a wind turbine provided in an embodiment of the present application;

[0096] Figure 2 A flow chart of a bearing detection method for a wind turbine in an actual application scenario provided by an embodiment of the present application;

[0097] Figure 3 A schematic diagram of a test result provided in an embodiment of the present application;

[0098] Figure 4 A schematic diagram of a test result provided in an embodiment of the present application;

[0099] Figure 5 A schematic diagram of a test result provided in an embodiment of the present application;

[0100] Figure 6 A schematic diagram of a test result provided in an embodiment of the present application;

[0101] Figure 7 A structural block diagram of a bearing detection device for a wind turbine provided in an embodiment of the present application. DETAILED DESCRIPTION

[0102] The embodiments of the present application are described below in conjunction with the accompanying drawings.

[0103] Wind turbines (also called wind turbines) can be roughly divided into two categories according to the mechanical drive method: direct drive and non-direct drive. Direct drive, as the name suggests, means that the impeller directly drives the generator to rotate and generate electricity. Non-direct drive units add a speed-increasing gearbox between the impeller and the generator. The purpose is to convert the low speed of the impeller into a speed suitable for the operation of the generator at the back end.

[0104] Direct-drive wind turbines use an outer rotor permanent magnet synchronous generator, with the impeller directly connected to the generator rotor, running at low speed, and no grid excitation is required. Direct-drive wind turbines include wind turbine impellers, multi-pole low-speed synchronous generators, full-power converters, etc. The control system of the wind turbine mainly relies on the converter to adjust the output current of the generator, thereby achieving torque regulation. Since the wind turbine in this type of generator set is directly connected to the generator without using a gearbox, that is, the "direct drive" wind turbine power generation technology, it is called a direct-drive wind turbine. The generator adopts a single-bearing structure, which further simplifies the transmission chain of the unit, making the structure of the unit more compact and simple.

[0105] The gearbox is one of the main components of a non-direct-drive wind turbine generator set. Usually, the speed of the wind wheel is very low, far from the speed required by the generator to generate electricity. It must be achieved through the speed-increasing effect of the gearbox gear pair, so the gearbox is also called a speed-increasing box. According to the overall layout requirements of the unit, the transmission shaft (commonly known as the main shaft) directly connected to the wind wheel hub is sometimes integrated with the gearbox. There is also a structure in which the main shaft and the gearbox are arranged separately, and a clamping sleeve device or a coupling is used to connect them. In order to increase the braking capacity of the unit, a brake device is often set at the input or output end of the gearbox, and the blade tip brake (fixed pitch wind wheel) or variable pitch brake device is used to jointly brake the unit transmission system. It can be seen that whether in a direct-drive or non-direct-drive generator, bearings are components that play an important role in the operation of wind turbines.

[0106] Wind turbines are exposed to harsh environments such as humidity, corrosion, and strong winds for a long time. The designed service life of a wind turbine is usually up to several decades, but it is unknown whether the generator can work safely and stably under high-intensity load conditions. Therefore, the bearing wear of direct-drive and non-direct-drive wind turbines is a problem that is likely to occur as the operating time increases. Therefore, in order to ensure the stable operation of wind turbines, it is necessary to accurately and effectively detect the bearings of wind turbines. In the related art, the detection of the bearings of wind turbines requires relevant technicians to enter the wind turbine to observe the cracks on the bearings, or collect the grease around the bearings to analyze the iron content therein, which requires a lot of time and effort, and it is difficult to obtain more accurate detection results. The detection results cannot reflect the abnormal conditions of the wind turbine in a specific wind turbine generator set.

[0107] In order to solve the above technical problems, the present application provides a bearing wear detection method for wind turbines. The processing equipment can analyze the collected bearing temperature data of the target wind turbine generator set according to multiple unsupervised models with multiple abnormal data detection methods, so as to more accurately determine the wind turbine with abnormal bearing temperature data from a specific wind turbine generator set, thereby improving the accuracy of detecting the bearing status of the wind turbine generator.

[0108] It is understandable that the method can be applied to a processing device, which is a processing device capable of bearing detection, for example, a terminal device or a server with a bearing detection function. The method can be executed independently by a terminal device or a server, or it can be applied to a network scenario in which a terminal device and a server communicate, and is executed by the cooperation of a terminal device and a server. Among them, the terminal device can be a computer, a mobile phone and other devices. The server can be understood as an application server or a Web server. In actual deployment, the server can be an independent server or a cluster server.

[0109] Next, a bearing detection method for a wind turbine provided in an embodiment of the present application will be introduced in conjunction with the accompanying drawings.

[0110] See also Figure 1 , Figure 1 A flow chart of a bearing detection method for a wind turbine provided in an embodiment of the present application, the method comprising:

[0111] S101: Acquire bearing temperature data corresponding to a target wind turbine generator set.

[0112] It is understandable that during the operation of a wind turbine, the bearing will generate a certain amount of heat due to the friction between the impeller and the generator rotor, and when the bearing state is different, the generated bearing temperature data will also be different. For example, when the speed difference between the front and rear ends of the bearing is large, the temperature difference between the front and rear ends of the bearing will also be large. It can be seen that by analyzing the bearing temperature data, it is possible to detect whether the bearing state is abnormal to a certain extent.

[0113] The target wind turbine generator set is the wind turbine generator set targeted by the bearing test, and the target wind turbine generator set includes at least two wind turbines; the bearing temperature data is used to reflect the bearing temperature of each wind turbine in the target wind turbine generator set during operation. It is worth noting that since the bearing only generates heat when the wind turbine is running, in order to improve accuracy, the bearing temperature data collected in this application is in the operating state.

[0114] S102: Determine, by using multiple unsupervised models and based on bearing temperature data, a first abnormal probability of each generator in the target wind turbine generator set in a bearing temperature dimension.

[0115] In order to analyze the bearing temperature data more efficiently and conveniently, in the embodiment of the present application, the processing device can analyze the bearing temperature data through an unsupervised model. The advantage of the unsupervised model is that the processing device does not need to add abnormal or normal labels to the bearing temperature data. The unsupervised model can analyze the data by itself and determine the data standard of the group of data under normal circumstances, so as to automatically analyze and obtain the abnormal data based on the data standard.

[0116] Furthermore, in order to improve the accuracy of the data analysis results, the processing device can obtain multiple unsupervised models corresponding to multiple abnormal data detection methods, and jointly analyze the bearing temperature data through multiple unsupervised models, so as to obtain detection results under multiple detection methods. The processing device can comprehensively determine the first abnormal probability of the bearing temperature data, thereby making the obtained abnormal probability more comprehensive and accurate.

[0117] For example, in a possible implementation, multiple unsupervised models can be feature bagging models (FeatureBagging), isolation forest models (Isolation Forest), nearest neighbor node models (k-NearestNeighbor, referred to as KNN), average nearest neighbor node models (Average KNN). Among them, the abnormal data detection methods corresponding to these four unsupervised models are different. For example, in the isolation forest model, the abnormal data is mainly judged by analyzing the outliers in the data; in the nearest neighbor node model, if most of the K most similar samples (i.e., the nearest in the feature space) of a sample in the feature space belong to a certain category, then the sample also belongs to this category. This method only determines the category of the sample to be classified based on the category of the nearest one or several samples in the classification decision, so that the abnormal data with a large gap from the normal category can be analyzed. It can be understood that there may be some models in these multiple models that use similar or identical abnormal data detection methods. In this application, there is no need to ensure that each unsupervised model corresponds to a different detection method.

[0118] S103: Determine a wind turbine generator having a first abnormality probability greater than a first probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0119] Through multiple unsupervised models, the processing device can determine the probability of abnormal bearing temperature data of each wind turbine in the target wind turbine generator set. The processing device can preset a first probability threshold and determine the wind turbine with an abnormal probability greater than the first probability threshold as a wind turbine with abnormal bearing in the target wind turbine generator set.

[0120] For example, the processing device may set the first probability threshold to 0.4, as shown in the following table:

[0121]

[0122]

[0123] Among them, the data predicted as 0 is normal data, and the data predicted as 1 is abnormal data. It can be seen that the two wind turbines in the table have an abnormal probability greater than 0.4 in a certain unsupervised model, so they can both be regarded as wind turbines with abnormal bearings.

[0124] It can be seen from the above technical solution that when detecting the bearing, the bearing temperature data corresponding to the target wind turbine generator set can be obtained first, and the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation. Through multiple unsupervised models corresponding to multiple abnormal data detection methods, the bearing temperature data can be analyzed from multiple angles, so as to more accurately obtain the normal state of the bearing temperature of the target wind turbine generator set during operation, and then the more abnormal temperature data can be determined from the bearing temperature data. The abnormal bearing temperature data can reflect the probability of abnormality of the wind turbine generator bearing. By determining the wind turbine generator with a first abnormal probability greater than the first probability threshold as a bearing abnormal wind turbine generator in the target wind turbine generator set, accurate detection of wind turbine generators with abnormal bearings can be achieved without manual entry into the wind turbine generator, thereby improving the efficiency and accuracy of bearing detection and reducing the consumption of time and energy.

[0125] It is understandable that during the operation of a wind turbine, abnormal conditions that are not caused by bearing abnormalities may occur. For example, the wind turbine cannot operate in a good state when the wind speed is too low. At the same time, due to abnormalities in the data acquisition process, abnormal bearing temperature data may also appear in a normal bearing. For example, when collecting bearing temperature data through a temperature sensor, abnormal temperature data may be obtained due to abnormalities in the temperature sensor itself.

[0126] Based on this, in a possible implementation method, in order to make the bearing temperature data used for analysis more accurate, the processing device can also obtain the operating data corresponding to the target wind turbine generator set within a preset time period, and the operating data includes the pending bearing temperature data corresponding to the target wind turbine generator set within the preset time period, and associated data associated with the pending bearing temperature data, and the associated data includes any one or more combinations of wind turbine status data, active power data, wind turbine speed data, cabin temperature data, and cabin vibration data. Through these associated data, the processing device can analyze the operating status of the wind turbine corresponding to the collected bearing temperature data, so as to determine the bearing temperature data under normal operating conditions.

[0127] The processing device may first preset a data volume threshold and a generator quantity threshold. The data volume threshold is used to determine whether the amount of operating data is sufficient. The generator quantity threshold is used to determine whether the number of wind turbines in the target wind turbine generator set is sufficient, thereby ensuring that the unsupervised model has sufficient data for analysis and obtaining a more accurate abnormality detection result. In response to the data volume of the operating data being greater than the data volume threshold, and the number of wind turbines in the target wind turbine generator set being greater than the generator quantity threshold, the processing device may filter the operating data based on the associated data and the data filtering threshold. The data filtering threshold is used to reflect the normal operating state of the wind turbines in the target wind turbine generator set. Thus, the processing device may filter out the data in the operating data when the wind turbine generator is in a normal operating state, and to a certain extent eliminate the interference caused by other abnormal factors to the bearing detection. The processing device may determine the pending bearing temperature data included in the filtered operating data as the bearing temperature data corresponding to the target wind turbine generator set.

[0128] For example, the processing device may obtain the real-time operating data of the target wind turbine generator set for three days. The following table shows the contents of the real-time operating data:

[0129] Field English name Field Chinese name wtid Unit No. WTUR_TurSt_Rs_S Fan status WTUR_WSpd_Ra_F32 Wind speed rectime time WTUR_PwrAt_Ra_F32 Generator speed WGEN_Spd_Ra_F32 Active Power WTUR_Acce_Ra_F32_x X-direction vibration of the cabin WTUR_Acce_Ra_F32_y Cabin Y-direction vibration WNAC_Temp_Ra_F32 Cabin temperature WGEN_Temp_Ra_F32_BackAxis1 Generator rear bearing temperature 1 WGEN_Temp_Ra_F32_BackAxis2 Generator rear bearing temperature 2 WGEN_Temp_Ra_F32_FrontAxis1 Generator front bearing temperature 1 WGEN_Temp_Ra_F32_FrontAxis2 Generator front bearing temperature 2

[0130] Subsequently, the processing device can determine whether the total amount of data is greater than 30,000 and whether the number of wind turbines is greater than 3. If not, the bearing abnormality detection is terminated. Subsequently, the processing device can convert the time column to the time type and convert other columns to the digital type to facilitate input into the unsupervised model for analysis. At the same time, the processing device can determine whether there are all empty columns. If so, it is determined to be invalid data and no analysis is performed. The processing device can also check whether there is empty data. If so, the entire row of data is deleted to ensure the validity of the data.

[0131] The processing device may determine the following conditions as data filtering thresholds:

[0132] The fan status data is 5;

[0133] Active power data is greater than 50;

[0134] The fan speed data is greater than 8;

[0135] The cabin temperature data is greater than -15 degrees and less than 50 degrees;

[0136] The cabin vibration is not 0 on both the X and Y axes.

[0137] In addition to temperature data that can reflect the operating status of the bearing, the processing device can further integrate operating data from other dimensions to jointly analyze the bearing status. For example, during the operation of a wind turbine, the rotation of the bearing will cause the vibration of the wind turbine nacelle. Therefore, when the rotation of the bearing is abnormal, the vibration of the wind turbine nacelle may also be different from that in the normal state. Therefore, in one possible implementation, the processing device can combine the bearing temperature data and the nacelle vibration data to jointly detect the bearing status.

[0138] In this embodiment, the operating data may include cabin vibration data, and the processing device may determine the cabin vibration data associated with the bearing temperature data as the target cabin vibration data corresponding to the target wind turbine generator set. The target cabin data is data reflecting the bearing status of the same operating process as the bearing temperature data.

[0139] The processing device can determine the second abnormality probability of each generator in the target wind turbine generator set in the bearing vibration dimension according to the target cabin vibration data through multiple unsupervised models. When determining a wind turbine generator with an abnormal bearing, the processing device can determine a wind turbine generator with a first abnormality probability greater than a first probability threshold and a second abnormality probability greater than a second probability threshold as a wind turbine generator with an abnormal bearing in the target wind turbine generator set. For example, the processing device can obtain the first abnormality probability and the second abnormality probability output by multiple unsupervised models respectively, and then determine a wind turbine generator with a first abnormality probability greater than the first probability threshold obtained in at least one unsupervised model and a second abnormality probability greater than the second probability threshold obtained in at least one unsupervised model as a wind turbine generator with an abnormal bearing in the target wind turbine generator set. For example, in combination with the above table and the table below,

[0140]

[0141] It can be seen from the table that when the first probability threshold and the second probability threshold are both 0.45, the two wind turbines numbered 411725010 and 411725011 in the table have at least one unsupervised model output abnormal probability greater than 0.45 in the bearing vibration dimension detection and the bearing temperature dimension detection. Therefore, they can be determined as wind turbines with abnormal bearings in the target wind turbine group.

[0142] In addition, in order to perform more detailed inspection of the bearings, in a possible implementation, the processing device can also classify the operating state of the wind turbine based on the rotational speed of the wind turbine, and determine the abnormal probability corresponding to the wind turbine in each operating state.

[0143] In this embodiment, the operation data may include wind turbine speed data, and the processing device may determine the wind turbine speed data associated with the bearing temperature data as the target wind turbine speed data corresponding to the target wind turbine generator set, and the target wind turbine speed data is the speed data reflecting the same wind turbine operation process as the bearing temperature data. The processing device may group the bearing temperature data and the target cabin vibration data according to the target wind turbine speed data, so as to obtain the bearing temperature data and cabin vibration data of the wind turbine generator under different operating conditions. The processing device may determine, in at least one speed grouping, a wind turbine corresponding to at least one of the first abnormal probability and the second abnormal probability greater than the third probability threshold as a bearing abnormal wind turbine in the target wind turbine generator set, so as to accurately detect the bearings with a higher abnormal probability in one or more operating conditions, thereby improving the comprehensiveness and accuracy of the detection.

[0144] Specifically, when the bearing is detected based on the bearing temperature data, the processing device can use different data processing methods for detection for different wind turbine models. For example, in one possible implementation, the wind turbine in the target wind turbine generator set can be a direct-drive wind turbine, and the bearing temperature data has associated cabin temperature data. Since in a direct-drive wind turbine, the bearing is responsible for connecting the impeller and generator of the wind turbine, the processing device can install two temperature sensors on the impeller side and the generator side to collect the bearing temperature data.

[0145] The bearing temperature data may include first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data, and second bearing generator side temperature data. The processing device may determine the first temperature data MAX_TEMP according to the average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0146] Determine the second temperature data MAX_NACELL_DIFF according to the difference between the maximum value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data;

[0147] Determine the third temperature data MEAN_FRONT_TEMP according to the average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data;

[0148] Determine fourth temperature data MEAN_BACK_TEMP according to the average value of the first bearing generator side temperature data and the second bearing generator side temperature data;

[0149] Determine fifth temperature data DIFF_FB_TEMP according to the difference between the third temperature data and the fourth temperature data;

[0150] Determine sixth temperature data STD_TEMP according to the standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0151] According to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data, a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator set is determined. For example, the six temperature data can be converted from 0 to 1 to form a matrix. The first temperature data matrix can reflect the temperature conditions on the bearing from multiple angles. The processing device can determine the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension according to the first temperature data matrix through multiple unsupervised models.

[0152] If the wind turbine in the target wind turbine generator set is a non-direct drive wind turbine, the bearing is connected to the high-speed shaft of the gearbox, and the bearing temperature data has associated cabin temperature data, and the bearing temperature data includes the gearbox high-speed shaft front shaft temperature data and the gearbox high-speed shaft rear shaft temperature data. When performing bearing detection, the processing device can determine the seventh temperature data BFGEARBEAR_TEMP_DIFF according to the difference between the gearbox high-speed shaft front shaft temperature data and the gearbox high-speed shaft rear shaft temperature data;

[0153] Determine the eighth temperature data BGEARBEAR_NACELL_DIFF according to the difference between the rear axle temperature data of the high-speed shaft of the gearbox and the cabin temperature data;

[0154] The gearbox high-speed shaft rear axle temperature data is determined as the ninth temperature data BGEARBEAR_TEMP;

[0155] The processing device can determine the second temperature data matrix corresponding to each wind turbine in the target wind turbine generator group based on the seventh temperature data, the eighth temperature data and the ninth temperature data, and then determine the first abnormality probability of each generator in the target wind turbine generator group in the bearing temperature dimension based on the second temperature data matrix through multiple unsupervised models.

[0156] In addition to temperature data, the processing device can also process cabin vibration data in a targeted manner. For example, in a possible implementation, the target cabin vibration data can include cabin transverse axis vibration data and cabin longitudinal axis vibration data, and the processing device can determine the first vibration data ACCE_X_M based on the difference between the absolute value of the cabin transverse axis vibration data and the absolute value of the mean value of the cabin transverse axis vibration data;

[0157] The second vibration data ACCE_Y_M is determined based on the difference between the absolute value of the cabin longitudinal axis vibration data and the absolute value of the mean of the cabin longitudinal axis vibration data, and then the vibration data matrix corresponding to each wind turbine in the target wind turbine generator set is determined based on the first vibration data and the second vibration data. Through multiple unsupervised models, the second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension can be determined based on the vibration data matrix.

[0158] In one possible implementation, the wind turbines in the target wind turbine generator group may be non-direct drive wind turbines. In this case, the processing device may also determine the third vibration data ACCE_MAX_M based on the maximum value of the cabin transverse axis vibration data and the cabin longitudinal axis vibration data, and then determine the vibration data matrix corresponding to each wind turbine in the target wind turbine generator group based on the first vibration data, the second vibration data and the third vibration data.

[0159] In the above detection process, the processing device may also appropriately adjust the first probability threshold and the second probability threshold according to the result of the bearing abnormality detection. For example, when the processing device finds that there are many wind turbines with abnormal bearing detection results, but there are not many abnormal wind turbines in fact, the abnormal probability threshold may be appropriately increased.

[0160] In order to facilitate understanding of the technical solution provided in the embodiment of the present application, a bearing detection method for a wind turbine provided in the embodiment of the present application will be introduced below in combination with an actual application scenario.

[0161] See also Figure 2 , Figure 2 A flow chart of a bearing detection method for a wind turbine in an actual application scenario provided in an embodiment of the present application, the method comprising:

[0162] S201: Read the real-time operating data of all units in an electric field for 3 days.

[0163] S202: Determine whether the amount of unit operation data is large

[0164] The number of wind turbines is greater than 30,000.

[0165] If yes, continue the test; if no, end.

[0166] S203: The read data is converted into digital type, and the time is forcibly converted into time type.

[0167] S204: Filtering is performed based on a data filtering threshold.

[0168] S205: Process the bearing temperature data to obtain six temperature data.

[0169] S206: Process the cabin vibration data to obtain three vibration data.

[0170] S207: Determine a temperature data matrix according to the six temperature data.

[0171] S208: Determine a vibration data matrix according to the three vibration data.

[0172] S209: Divide each unit into bins according to each speed range, and the amount of data 0 or 1 in each bin is greater than 1000.

[0173] If grouping can be performed based on this condition, continue the test; otherwise, end the test.

[0174] S210: Detection condition: 1. The first abnormal probability and the second abnormal probability are both greater than 0.4 in at least one model.

[0175] 2. The first abnormality probability and the second abnormality probability exist in at least one speed range and are greater than 0.9.

[0176] If there is a wind turbine generator that meets the above regulation, jump to S211.

[0177] S211: Warning generator bearing abnormality

[0178] As shown in the figure, Figure 3 and Figure 4 Graphical display of data of target wind turbines, Figure 5 and Figure 6 The graphical display of the data of the surrounding units in the same period shows that the temperature difference of the wind turbines in the wind turbine generator set should be small and the vibration data should be relatively concentrated. However, the temperature and vibration of the wind turbines in the target wind turbine generator set are discrete and offset.

[0179] Based on the bearing detection method for a wind turbine provided in the above embodiment, the present application also provides a bearing detection device for a wind turbine, see Figure 7 , Figure 7 This is a structural block diagram of a bearing detection device 700 for a wind turbine provided in an embodiment of the present application. The device includes an acquisition unit 701, a first determination unit 702, and a second determination unit 703:

[0180] An acquisition unit 701 is used to acquire bearing temperature data corresponding to a target wind turbine generator set, wherein the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation;

[0181] A first determination unit 702 is used to determine a first abnormal probability of each generator in the target wind turbine generator set in a bearing temperature dimension according to the bearing temperature data by using a plurality of unsupervised models, wherein the plurality of unsupervised models correspond to a plurality of abnormal data detection methods;

[0182] The second determining unit 703 is configured to determine the wind turbine generator having the first abnormality probability greater than the first probability threshold as the wind turbine generator having the bearing abnormality in the target wind turbine generator group.

[0183] In a possible implementation, the acquiring unit 701 is specifically configured to:

[0184] Acquiring operation data corresponding to the target wind turbine generator set within a preset period of time, wherein the operation data includes pending bearing temperature data corresponding to the target wind turbine generator set within the preset period of time, and associated data associated with the pending bearing temperature data, wherein the associated data includes any one or more combinations of wind turbine status data, active power data, wind turbine speed data, cabin temperature data, and cabin vibration data;

[0185] In response to the data volume of the operating data being greater than a data volume threshold and the number of wind turbines in the target wind turbine generator set being greater than a turbine number threshold, filtering the operating data based on the associated data and a data filtering threshold, wherein the data filtering threshold is used to reflect the normal operating status of the wind turbines in the target wind turbine generator set;

[0186] The undetermined bearing temperature data included in the screened operating data is determined as the bearing temperature data corresponding to the target wind turbine generator set.

[0187] In a possible implementation, the operation data includes the cabin vibration data, and the apparatus 700 further includes a third determining unit and a fourth determining unit:

[0188] A third determining unit, configured to determine the nacelle vibration data associated with the bearing temperature data as target nacelle vibration data corresponding to the target wind turbine generator set;

[0189] a fourth determination unit, configured to determine, by using the multiple unsupervised models and according to the target nacelle vibration data, a second abnormal probability of each generator in the target wind turbine generator set in a bearing vibration dimension;

[0190] The second determining unit is specifically configured to:

[0191] A wind turbine generator having a first abnormal probability greater than a first probability threshold and a second abnormal probability greater than a second probability threshold is determined as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

[0192] In a possible implementation, the operation data includes the fan speed data, and the device 700 further includes a fifth determination unit, a grouping unit, and a sixth determination unit:

[0193] a fifth determining unit, configured to determine the wind turbine speed data associated with the bearing temperature data as the target wind turbine speed data corresponding to the target wind turbine generator set;

[0194] A grouping unit, configured to group the bearing temperature data and the target nacelle vibration data according to the target fan speed data;

[0195] The sixth determination unit is used to determine, in at least one speed group, a wind turbine corresponding to at least one of the first abnormal probability and the second abnormal probability greater than a third probability threshold as a wind turbine with a bearing abnormality in the target wind turbine group.

[0196] In a possible implementation manner, the second determining unit is specifically configured to:

[0197] A wind turbine having a first abnormality probability obtained in at least one unsupervised model greater than a first probability threshold and a second abnormality probability obtained in at least one unsupervised model greater than a second probability threshold is determined as a wind turbine having a bearing abnormality in the target wind turbine group.

[0198] In a possible implementation, the wind turbine in the target wind turbine generator set is a direct-drive wind turbine, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data, and second bearing generator side temperature data, and the first determining unit 702 is specifically used to:

[0199] Determine first temperature data according to an average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0200] determining second temperature data according to a difference between a maximum value among the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data;

[0201] Determine third temperature data according to an average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data;

[0202] Determine fourth temperature data according to an average value of the first bearing generator side temperature data and the second bearing generator side temperature data;

[0203] determining fifth temperature data according to a difference between the third temperature data and the fourth temperature data;

[0204] Determine sixth temperature data according to a standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data;

[0205] Determine a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data;

[0206] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the first temperature data matrix through the multiple unsupervised models.

[0207] In a possible implementation, the wind turbine in the target wind turbine generator set is a non-direct drive wind turbine, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes gearbox high-speed shaft front shaft temperature data and gearbox high-speed shaft rear shaft temperature data, and the first determining unit 702 is specifically used to:

[0208] Determine seventh temperature data according to the difference between the temperature data of the front axle of the gearbox high-speed shaft and the temperature data of the rear axle of the gearbox high-speed shaft;

[0209] Determine eighth temperature data according to the difference between the gearbox high-speed shaft rear axle temperature data and the cabin temperature data;

[0210] Determine the gearbox high-speed shaft rear axle temperature data as ninth temperature data;

[0211] Determine a second temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the seventh temperature data, the eighth temperature data and the ninth temperature data;

[0212] The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the second temperature data matrix through the multiple unsupervised models.

[0213] In a possible implementation, the target cabin vibration data includes cabin transverse axis vibration data and cabin longitudinal axis vibration data, and the fourth determination unit is specifically configured to:

[0214] determining first vibration data according to a difference between an absolute value of the cabin transverse axis vibration data and an absolute value of a mean value of the cabin transverse axis vibration data;

[0215] determining second vibration data according to a difference between an absolute value of the cabin longitudinal axis vibration data and an absolute value of a mean value of the cabin longitudinal axis vibration data;

[0216] Determine a vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data;

[0217] The second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension is determined according to the vibration data matrix through the multiple unsupervised models.

[0218] In a possible implementation, the wind turbine in the target wind turbine generator group is a non-direct drive wind turbine, and the method further includes a seventh determining unit:

[0219] a seventh determination unit, configured to determine third vibration data according to a maximum value of the cabin transverse axis vibration data and the cabin longitudinal axis vibration data;

[0220] The fourth determining unit is specifically configured to:

[0221] A vibration data matrix corresponding to each wind turbine in the target wind turbine generator group is determined according to the first vibration data, the second vibration data and the third vibration data.

[0222] A person of ordinary skill in the art can understand that all or part of the steps of implementing the above-mentioned method embodiment can be completed by hardware related to program instructions, and the above-mentioned program can be stored in a computer-readable storage medium. When the program is executed, it executes the steps of the above-mentioned method embodiment; and the above-mentioned storage medium can be at least one of the following media: read-only memory (English: read-only memory, abbreviated: ROM), RAM, magnetic disk or optical disk, etc. Various media that can store program codes.

[0223] It should be noted that each embodiment in this specification is described in a progressive manner, and the same and similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device and system embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiments. The device and system embodiments described above are merely schematic, in which the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative work.

[0224] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A bearing detection method for a wind turbine generator, characterized in that: The method comprises: Acquire bearing temperature data corresponding to the target wind turbine generator set, wherein the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation; Determining, by using a plurality of unsupervised models, a first abnormal probability of each generator in the target wind turbine generator set in a bearing temperature dimension according to the bearing temperature data, wherein the plurality of unsupervised models correspond to a plurality of abnormal data detection methods; Determine a wind turbine generator having a first abnormal probability greater than a first probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group; The wind turbine generator in the target wind turbine generator set includes a direct-drive wind turbine generator, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data, and second bearing generator side temperature data. For the direct-drive wind turbine generator, the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the bearing temperature data by using multiple unsupervised models, including: Determine first temperature data according to an average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data; determining second temperature data according to a difference between a maximum value among the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data; Determine third temperature data according to an average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data; Determine fourth temperature data according to an average value of the first bearing generator side temperature data and the second bearing generator side temperature data; determining fifth temperature data according to a difference between the third temperature data and the fourth temperature data; Determine sixth temperature data according to a standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data; Determine a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data; The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the first temperature data matrix through the multiple unsupervised models.

2. The method according to claim 1, characterized in that The step of obtaining bearing temperature data corresponding to the target wind turbine generator set includes: Acquiring operation data corresponding to the target wind turbine generator set within a preset period of time, wherein the operation data includes pending bearing temperature data corresponding to the target wind turbine generator set within the preset period of time, and associated data associated with the pending bearing temperature data, wherein the associated data includes any one or more combinations of wind turbine status data, active power data, wind turbine speed data, cabin temperature data, and cabin vibration data; In response to the data volume of the operating data being greater than a data volume threshold and the number of wind turbines in the target wind turbine generator set being greater than a turbine number threshold, filtering the operating data based on the associated data and a data filtering threshold, wherein the data filtering threshold is used to reflect the normal operating status of the wind turbines in the target wind turbine generator set; The undetermined bearing temperature data included in the screened operating data is determined as the bearing temperature data corresponding to the target wind turbine generator set.

3. The method according to claim 2, characterized in that The operation data includes the cabin vibration data, and the method further includes: Determining the nacelle vibration data associated with the bearing temperature data as the target nacelle vibration data corresponding to the target wind turbine generator set; Determining, by means of the multiple unsupervised models and based on the target nacelle vibration data, a second abnormal probability of each generator in the target wind turbine generator set in a bearing vibration dimension; The step of determining the wind turbine generator having a first abnormal probability greater than a first probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group includes: A wind turbine generator having a first abnormal probability greater than a first probability threshold and a second abnormal probability greater than a second probability threshold is determined as a wind turbine generator having a bearing abnormality in the target wind turbine generator group.

4. The method according to claim 3, characterized in that The operation data includes the fan speed data, and the method further includes: Determining the wind turbine speed data associated with the bearing temperature data as the target wind turbine speed data corresponding to the target wind turbine generator set; grouping the bearing temperature data and the target nacelle vibration data according to the target fan speed data; A wind turbine generator corresponding to at least one of the first abnormal probability and the second abnormal probability greater than a third probability threshold in at least one speed group is determined as a wind turbine generator generator with a bearing abnormality in the target wind turbine generator group.

5. The method according to claim 3, characterized in that: The step of determining a wind turbine generator having a first abnormality probability greater than a first probability threshold and a second abnormality probability greater than a second probability threshold as a wind turbine generator having a bearing abnormality in the target wind turbine generator group includes: A wind turbine having a first abnormality probability obtained in at least one unsupervised model greater than a first probability threshold and a second abnormality probability obtained in at least one unsupervised model greater than a second probability threshold is determined as a wind turbine having a bearing abnormality in the target wind turbine group.

6. The method according to claim 1, characterized in that The wind turbine generators in the target wind turbine generator set include non-direct-drive wind turbine generators, the bearing temperature data have associated nacelle temperature data, the bearing temperature data include gearbox high-speed shaft front shaft temperature data and gearbox high-speed shaft rear shaft temperature data, and for the non-direct-drive wind turbine generators, the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the bearing temperature data through multiple unsupervised models, including: Determine seventh temperature data according to the difference between the temperature data of the front axle of the gearbox high-speed shaft and the temperature data of the rear axle of the gearbox high-speed shaft; Determine eighth temperature data according to the difference between the gearbox high-speed shaft rear axle temperature data and the cabin temperature data; Determine the gearbox high-speed shaft rear axle temperature data as ninth temperature data; Determine a second temperature data matrix corresponding to each wind turbine in the target wind turbine generator set according to the seventh temperature data, the eighth temperature data and the ninth temperature data; The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the second temperature data matrix through the multiple unsupervised models.

7. The method according to claim 3, characterized in that The target nacelle vibration data includes nacelle transverse axis vibration data and nacelle longitudinal axis vibration data, and determining the second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension according to the target nacelle vibration data through the multiple unsupervised models includes: determining first vibration data according to a difference between an absolute value of the cabin transverse axis vibration data and an absolute value of a mean value of the cabin transverse axis vibration data; determining second vibration data according to a difference between an absolute value of the cabin longitudinal axis vibration data and an absolute value of a mean value of the cabin longitudinal axis vibration data; Determine a vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data; The second abnormal probability of each generator in the target wind turbine generator set in the bearing vibration dimension is determined according to the vibration data matrix through the multiple unsupervised models.

8. The method according to claim 7, characterized in that The wind turbine in the target wind turbine generator set is a non-direct drive wind turbine, and the method further includes: determining third vibration data according to a maximum value of the cabin transverse axis vibration data and the cabin longitudinal axis vibration data; Determining the vibration data matrix corresponding to each wind turbine in the target wind turbine generator set according to the first vibration data and the second vibration data includes: A vibration data matrix corresponding to each wind turbine in the target wind turbine generator group is determined according to the first vibration data, the second vibration data and the third vibration data.

9. A bearing detection device for a wind turbine, characterized in that: The device comprises an acquisition unit, a first determination unit and a second determination unit: The acquisition unit is used to acquire bearing temperature data corresponding to the target wind turbine generator set, wherein the bearing temperature data is used to reflect the bearing temperature of each wind turbine generator in the target wind turbine generator set during operation; The first determination unit is used to determine the first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension according to the bearing temperature data through multiple unsupervised models, and the multiple unsupervised models correspond to multiple abnormal data detection methods; The second determination unit is used to determine the wind turbine generator whose first abnormality probability is greater than the first probability threshold as the wind turbine generator with bearing abnormality in the target wind turbine generator group; The wind turbine in the target wind turbine generator set includes a direct-drive wind turbine, the bearing temperature data has associated nacelle temperature data, the bearing temperature data includes first bearing impeller side temperature data, second bearing impeller side temperature data, first bearing generator side temperature data and second bearing generator side temperature data, and for the direct-drive wind turbine, the first determining unit is specifically used to: Determine first temperature data according to an average value of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data; determining second temperature data according to a difference between a maximum value among the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data and the second bearing generator side temperature data and the nacelle temperature data associated with the bearing temperature data; Determine third temperature data according to an average value of the first bearing impeller side temperature data and the second bearing impeller side temperature data; Determine fourth temperature data according to an average value of the first bearing generator side temperature data and the second bearing generator side temperature data; determining fifth temperature data according to a difference between the third temperature data and the fourth temperature data; Determine sixth temperature data according to a standard deviation of the first bearing impeller side temperature data, the second bearing impeller side temperature data, the first bearing generator side temperature data, and the second bearing generator side temperature data; Determine a first temperature data matrix corresponding to each wind turbine in the target wind turbine generator group according to the first temperature data, the second temperature data, the third temperature data, the fourth temperature data, the fifth temperature data and the sixth temperature data; The first abnormal probability of each generator in the target wind turbine generator set in the bearing temperature dimension is determined according to the first temperature data matrix through the multiple unsupervised models.

Citation Information

Patent Citations

  • Wind generating set blade abnormal recognition method and device

    CN108087210A

  • Early warning method for abnormal temperature of main bearing of wind driven generator

    CN113312851A

  • Abnormality detection system and abnormality detection method of wind power generator

    JP2019027324A

Cited By

  • Fault detection device for wind driven generator

    CN120251462A