Wind turbine bearing sticking anomaly identification system, method, and computer equipment

Abstract

The present invention belongs to the technical field of wind turbine bearing abnormality identification, and relates to a system and method for identifying bearing jam anomalies of wind turbines. The method includes three parts: checking the abnormal rotation speed in the standby state; checking the abnormal startup time during the startup process; and abnormal bearing temperature rise in the power generation state. The three conditions are judged, and the bearing abnormality is located when all are met. The algorithm can be transplanted into the field controller to conduct regular inspections of all units in the field, and issue early warning information to prompt operation and maintenance personnel to conduct further inspections and take action in advance to avoid the expansion of bearing problems, extend the operating life of the unit bearings, and avoid power generation losses of the unit. The operation big data of the wind turbine is analyzed, and by comparing various anomalies in the standby process, startup process, and operation process, the bearing problems in the unit transmission chain are effectively identified.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind turbine bearing abnormality identification, and in particular relates to a wind turbine bearing sticking abnormality identification system, method and computer equipment. Background Art

[0002] The bearings in a wind turbine's drive train are core components. Due to differences in direct-drive, semi-direct-drive, and doubly-fed turbine structures, the number and structure of bearings in the drive train vary. Wind turbines operate outdoors year-round, under harsh operating conditions and present significant operational and maintenance challenges. In recent years, with the increasing size and cost reduction of wind turbines, bearing issues, particularly those in the main drive train, have become increasingly prominent.

[0003] If any bearing on the main transmission chain is damaged due to improper maintenance or load impact, it will affect the operation of the entire unit, causing the standby process to be prolonged and difficulty in meeting the starting conditions; there will be a jamming phenomenon during startup, and the startup speed will be slow; during operation, the bearing will heat up seriously and the power generation efficiency will be lower than that of other units. Over time, the bearing will deteriorate rapidly until the unit cannot operate normally.

[0004] There are many methods for bearing early warning in the wind power field. There are two main starting points. Both of these methods have their own shortcomings:

[0005] One method is based on vibration. For example, CN113295419A describes a method for warning faults in medium-speed bearings in wind turbine gearboxes. This involves installing a vibration sensor near the bearing to collect high-frequency vibration signals. Spectral analysis of characteristic frequency changes in the bearing vibration signals is then used to determine whether the bearing is abnormal. This method requires additional sensing equipment, and the data is difficult to store and obtain due to the high sampling frequency.

[0006] The second method is based on temperature. For example, CN110907066B describes a wind turbine gearbox bearing temperature status monitoring method based on a deep learning model. Using temperature data collected by temperature sensors placed on the bearings, a series of algorithms are designed to extract bearing temperature changes to determine if the bearings are abnormal. This method has a high false positive rate and uses temperature and temperature rise to determine bearing abnormalities. However, there are many causes for temperature anomalies, and bearing abnormalities are only one of them. Other factors, such as abnormal temperature sensor data collection and cooling system anomalies, can also lead to false positives, reducing the accuracy of early warnings. Summary of the Invention

[0007] The purpose of the present invention is to provide a system and method for identifying abnormal bearing jamming in a wind turbine generator set, which solves the problems of the current bearing warning methods in the wind power field that require the installation of additional equipment and have low warning accuracy.

[0008] The present invention is achieved through the following technical solutions:

[0009] A method for identifying abnormal bearing seizure in a wind turbine generator set comprises the following steps:

[0010] S1. Obtain the unit operation data of the target wind farm and perform data cleaning. The cleaned data includes the standby speed data of each unit, the startup duration data in the startup state, and the bearing temperature rise data;

[0011] S2. The standby speed data is processed by the standby speed model to output the average standby speed of each unit ωiavg and the average standby speed of all units ωavg;

[0012] The startup time data in the startup state is passed through the startup time model to output the average startup time of each unit τsiavg and the average startup time of all units τsavg;

[0013] The bearing temperature rise data is processed by the bearing temperature rise model to output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin;

[0014] Where i = 1, 2...N, i represents the unit number; k = 1, 2...K, k represents the number of compartments;

[0015] S3. Determine whether the wind turbine bearing has stuck abnormally based on the following three conditions:

[0016] Condition 1: Compare the average standby speed of all units, ωavg, with the average standby speed of each unit, ωiavg. If the speed exceeds the threshold, condition 1 is considered to be met.

[0017] Condition 2: Compare the average startup time of each unit, τsiavg, with the average startup time of all units, τsavg. If the time exceeds the threshold, condition 2 is considered to be met.

[0018] Condition 3: Compare the average bearing temperature rise (Tki) within each power bin of each unit with the average bearing temperature rise (Tkavg) of all units in each power bin. If the number of bins exceeding the temperature rise threshold exceeds the preset number of bins, condition 3 is considered met.

[0019] For a certain wind turbine generator set i, when all three conditions are met, it is determined that the bearing of the wind turbine generator set is stuck abnormally.

[0020] Furthermore, in S2, the standby speed data is processed by the standby speed model to output the average standby speed of each unit ωiavg and the average standby speed of all units ωavg; the specific calculation process of the standby speed model is:

[0021] 2.11) Identify the status of each unit's operating data and identify the data marked as standby;

[0022] 2.12) Calculate the average impeller speed ωij of all units in the site during each standby period, where i = 1, 2...N, i represents the unit number; j = 1, 2...M, j represents the number of standby periods;

[0023] 2.13) Sort the standby speeds ωij of all units on site, remove the x% maximum and x% minimum values, and obtain the filtered data;

[0024] 2.14) Calculate the average standby speed ωiavg of each unit based on the filtered data, and take the average standby speed ωavg of all units in the field;

[0025] Then take the average value of the filtered data to obtain the average standby speed ωavg of all units.

[0026] Furthermore, in S2, the startup duration data in the startup state is processed by the startup duration model to output the average startup duration τsiavg of each unit and the average startup duration τsavg of all units in the field; the specific calculation process of the startup duration model is:

[0027] 2.21) Identify the status of each unit's operating data and identify the data marked as startup;

[0028] Calculate the duration of each startup process τs ij for all units in the site, where i = 1, 2...N, i represents the unit number; j = 1, 2...M, j represents the number of startups;

[0029] 2.22) Sort the startup durations τs ij of all units on site, remove the x% maximum and x% minimum values, and obtain qualified startup duration data;

[0030] 2.23) Calculate the qualified startup duration data of each unit to obtain the average startup duration τsiavg of each unit;

[0031] Then take the average of the qualified startup time data to obtain the average startup time τsavg of all units.

[0032] Furthermore, in S2, the bearing temperature rise data is processed by the bearing temperature rise model to output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin. The calculation process of the bearing temperature rise model is as follows:

[0033] 2.31) Identify the power generation state, extract and calculate the temperature rise of the transmission chain bearings of all organic units in the site under the power generation state, and record the power generation power P corresponding to the temperature rise;

[0034] 2.32) Divide the temperature rise into bins based on the generated power. The power range is 0-Pr, where Pr is the rated power. Divide 0-Pr into k bins and calculate the average bearing temperature rise Tki in each power bin for each unit.

[0035] 2.33) Calculate the average temperature rise T1avg, T2avg, ..., Tkavg of all units in each power compartment;

[0036] Tkavg=(Tk1+Tk2+…+TkN) / N, where N is the unit number.

[0037] Furthermore, in S3, condition 1 is specifically:

[0038] Compare the average standby speed of all units ωavg with the average standby speed of each unit ωiavg. If it exceeds the speed threshold ω1, then condition 1 is considered to be met:

[0039] ωavg-ωiavg=ωDi>ω1.

[0040] Furthermore, in S3, condition 2 is specifically:

[0041] Compare the average startup time of each unit τsiavg with the average startup time of all units τsavg. If it exceeds a certain threshold τs1, it is considered that condition 2 is met:

[0042] τsiavg-τsavg=τsDi>τs1.

[0043] Furthermore, in S3, condition 3 is specifically:

[0044] If the difference between Tki and Tkavg exceeds the temperature rise threshold T0, and the number of compartments exceeding the temperature rise threshold T0 exceeds the preset number of compartments, it is determined that condition 3 is met.

[0045] Further, prompts are given based on the results of S3, specifically:

[0046] For a certain wind turbine generator set i, if the three conditions are met at the same time, the operation and maintenance prompt of the unit will be issued;

[0047] The threshold T0 in condition 3 can be set to three levels: T1, T2, and T3, with T3>T2>T1:

[0048] When T0 is selected as T1, the bearing abnormality will be reported;

[0049] When T0 is selected as T2, a bearing abnormality warning is issued;

[0050] When T0 is selected as T3, the bearing abnormality is reported urgently.

[0051] A wind turbine bearing sticking anomaly identification system, comprising:

[0052] The data processing module is used to clean the operating data of the units of the target wind farm. The cleaned data includes the standby speed data of each unit, the startup time data in the startup state, and the bearing temperature rise data;

[0053] The standby speed model is used to process the standby speed data and output the average standby speed ωiavg of each unit and the average standby speed ωavg of all units in the field;

[0054] The startup duration model is used to process the startup duration data in the startup state and output the average startup duration τsiavg of each unit and the average startup duration τsavg of all units in the field;

[0055] The bearing temperature rise model is used to process the bearing temperature rise data and output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin;

[0056] Where i = 1, 2...N, i represents the unit number; k = 1, 2...K, k represents the number of compartments;

[0057] S3, abnormal identification module, is used to determine whether there is any abnormality in the bearing of the wind turbine generator set:

[0058] Condition 1: Compare the average standby speed of all units, ωavg, with the average standby speed of each unit, ωiavg. If the speed exceeds the threshold, condition 1 is considered to be met.

[0059] Condition 2: Compare the average startup time of each unit, τsiavg, with the average startup time of all units, τsavg. If the time exceeds the threshold, condition 2 is considered to be met.

[0060] Condition 3: Compare the average bearing temperature rise (Tki) within each power bin of each unit with the average bearing temperature rise (Tkavg) of all units in each power bin. If the number of bins exceeding the temperature rise threshold exceeds the preset number of bins, condition 3 is considered met.

[0061] When all three conditions are met for a certain wind turbine generator set i, it is determined that the bearing of the wind turbine generator set of the set is stuck abnormally.

[0062] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the method for identifying abnormal bearing sticking of a wind turbine generator set are implemented.

[0063] Compared with the prior art, the present invention has the following beneficial technical effects:

[0064] The present invention discloses a method for identifying abnormal bearing jamming in a wind turbine generator set. The algorithm includes three parts: 1. Checking for abnormal rotational speed in the standby state; 2. Checking for abnormal startup time during the startup process; 3. Checking for abnormal bearing temperature rise in the power generation state. The three conditions are judged, and when all are met, the bearing abnormality is located. The operation big data of the wind turbine is analyzed, and by comparing various abnormalities during the standby process, startup process, and operation process, the bearing problems in the unit's transmission chain are effectively identified. When a fault is identified, an early warning is issued, and on-site personnel are prompted to perform predictive maintenance for early problems, such as extending the service life of the bearing by controlling load reduction and increasing lubrication effect, ensuring the stable operation of the unit. For bearings that have been severely deteriorated, a replacement plan can be implemented according to meteorological information to avoid unit failure and shutdown during the strong wind season and reduce power generation losses. Through the big data analysis method, a longitudinal comparison of the entire wind turbine over a period of operation is carried out to effectively identify abnormal bearing problems in the unit's transmission chain.

[0065] Furthermore, this method can be transplanted into computer equipment to conduct regular inspections of all units in the field and issue early warning information to prompt operation and maintenance personnel to conduct further inspections and take action in advance to avoid the expansion of bearing problems, extend the operating life of unit bearings, and avoid unit power generation losses. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] Figure 1 The present invention is a flow chart of a method for identifying abnormal bearing seizure in a wind turbine generator set. DETAILED DESCRIPTION

[0067] In order to make the purpose, technical solutions and advantages of the present invention more clear, the following is a further detailed description with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. That is, the embodiments described are only part of the embodiments of the present invention, not all of the embodiments.

[0068] The components described and illustrated in the drawings and embodiments of the present invention may be arranged and designed in a variety of different configurations. Therefore, the detailed description of the embodiments of the present invention provided in the following drawings is not intended to limit the scope of the claimed invention, but merely represents a selected embodiment of the present invention. All other embodiments derived by those skilled in the art based on the drawings and embodiments of the present invention without inventive effort shall fall within the scope of protection of the present invention.

[0069] It should be noted that the terms "comprises", "includes" or any other variations are intended to cover non-exclusive inclusion, so that a process, element, method, article or apparatus that includes a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to the process, element, method, article or apparatus.

[0070] The present invention discloses a method for identifying abnormal bearing seizure in a wind turbine generator set, comprising the following steps:

[0071] S1. Obtain the unit operation data of the target wind farm and perform data cleaning. The cleaned data includes the standby speed data of each unit, the startup duration data in the startup state, and the bearing temperature rise data;

[0072] S2. The standby speed data is processed by the standby speed model to output the average standby speed of each unit ωiavg and the average standby speed of all units ωavg;

[0073] The startup time data in the startup state is passed through the startup time model to output the average startup time of each unit τsiavg and the average startup time of all units τsavg;

[0074] The bearing temperature rise data is processed by the bearing temperature rise model to output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin;

[0075] Where i = 1, 2...N, i represents the unit number; k = 1, 2...K, k represents the number of compartments;

[0076] S3. Determine whether the wind turbine bearing has stuck abnormally based on the following three conditions:

[0077] Condition 1: Compare the average standby speed of all units, ωavg, with the average standby speed of each unit, ωiavg. If the speed exceeds the threshold, condition 1 is considered to be met.

[0078] Condition 2: Compare the average startup time of each unit, τsiavg, with the average startup time of all units, τsavg. If the time exceeds the threshold, condition 2 is considered to be met.

[0079] Condition 3: Compare the average bearing temperature rise (Tki) within each power bin of each unit with the average bearing temperature rise (Tkavg) of all units in each power bin. If the number of bins exceeding the temperature rise threshold exceeds the preset number of bins, condition 3 is considered met.

[0080] When all three conditions of a certain wind turbine generator set are met, it is determined that the bearing of the wind turbine generator set is abnormally stuck.

[0081] The features and performance of the present invention are further described in detail below with reference to the embodiments.

[0082] like Figure 1 As shown, the present invention's method for identifying bearing jam anomalies in wind turbine generator sets consists of four main steps: data preparation, model calculation, anomaly identification, and operation and maintenance prompts. This process can be implemented in a field controller (i.e., a centralized control and maintenance center). Once a bearing jam is detected, on-site personnel are prompted to perform further inspection and maintenance.

[0083] 1. Data preparation

[0084] 1) Collect and store operating data for the target wind farm, requiring no less than one month of data to fully reflect various operating conditions and external wind conditions, making the data more effective;

[0085] 2) The returned data is placed on the field server according to the specified location and number for storage;

[0086] 3) Data cleaning work, including deleting duplicate values, supplementing missing values, and removing interrupted data.

[0087] 2. Model calculation

[0088] The model calculation includes three parts: the calculation of the speed in standby mode, the calculation of the startup time in the startup state, and the calculation of the bearing temperature rise.

[0089] 2.1 Calculation of standby speed

[0090] In standby mode, due to the low wind energy, the normal unit speed will be low, usually around 0.5rpm. However, in a bearing stuck unit, due to bearing problems, the wind cannot effectively drive the rotor, so the speed change is very small.

[0091] 1) Identify the standby status of each unit data. The operation data channel contains records of the standby status. Normal units are divided into several states: standby, start-up, power generation, shutdown, and maintenance. Data marked as standby needs to be identified;

[0092] 2) Calculate the average impeller speed ωij (i = 1, 2 ... N, i represents the unit number) during each standby period of all units in the site (j = 1, 2 ... M, j represents the number of standby periods. Note that the total number of standby periods for each unit is different).

[0093] For example, for unit 1, a total of 100 standby times are identified. The speed information during each standby time is extracted and averaged, that is, the average speed values of unit 1 during the 100 standby times are obtained, ω11, ω12, ω13…ω1 100, if unit 2# has 103 standby times, we get ω21, ω22, ω23...ω2 103 Similarly, we can get the average speed of all units in the site during each standby period. Assuming that the last unit has a standby number of 89 times and the total number of the 50 units is 5000 times, the total number of ωij is 100+103+…+89=5000 times.

[0094] 3) Sort the standby speeds ωij of all units on site, remove the x% maximum and x% minimum values, and convert them into ωij'.

[0095] Assume ω11=0.01rpm,ω12=0.3rpm,……ω1 100 =0.2rpm, ω21=0.8rpm, ω12=0.2rpm…Sort these 5000 ωij values, find the maximum and minimum values that account for 1% (here x% is assumed to be 1%) and filter them out. After filtering out, 4900 data remain.

[0096] This step can remove extreme abnormal data caused by data or debugging reasons to avoid affecting the average standby speed.

[0097] 4) Calculate the average standby speed ωiavg of each unit and the average standby speed ωavg of all units based on these 4900 data points:

[0098] Since abnormal data has been filtered out in step 3), the number of standby times of each unit has changed. For example, the average speed of each standby process of unit 1# is ω11, ω13…ω1 respectively. 100 , because ω12 is too large and has been screened out, the number of speed values of unit 1 during standby is less than 100, but only 99. Then take the average standby speed value for unit 1#, and get ω1avg=(ω11+ω13+

[0099] ω14……+ω1 100 ) / 99, and similarly, the average standby speeds of other units are obtained, ω2avg, ω3avg, ... ω50avg (50 values are obtained, each value represents the average standby speed value of a unit.

[0100] Then take the average value of these 4900 ωij to obtain the average standby speed ωavg of all units.

[0101] 2.2 Calculation of startup time

[0102] Due to bearing sticking during startup, the average startup time of the unit with the problem is longer than that of other units under the same startup logic.

[0103] 1) Identify the status of each unit's operating data and identify the data marked as startup; calculate the duration of each startup process τs ij (i = 1, 2...N, i represents the unit number) for all units in the site (j = 1, 2...M, j represents the number of startups, note that the total number of startups for each unit is different);

[0104] For example, for Unit 1, a total of 100 startups are identified. The startup time is calculated by subtracting the startup start time from the startup end time. This gives 100 duration values for Unit 1: τs 11, τs 12, τs 13, …, τs 1100. If Unit 2 has 103 startups, then τs 21, τs 22, τs 23, …, τs 2103 are obtained. Similarly, the startup durations of all units at the site can be calculated. Assuming the last unit has 89 startups, and the total number of startups for all 50 units is 5000, the total number of τs ij is 100 + 103 + … + 89 = 5000.

[0105] 2) Sort the startup durations τs ij of all units on site, remove the x% maximum and x% minimum values, and convert them into τs ij'.

[0106] Assume that τs 11 = 40s, τs 12 = 2000s, ... τs 1100 = 70s, τs 21 = 60s, and τs 12 = 80s. Sort these 5000 τs ij values, find the maximum and minimum values that account for 1% (here x is assumed to be 1%), and filter them out. After filtering, 4900 data points remain.

[0107] This step removes extreme abnormal data due to data reasons or debugging reasons to avoid affecting the average startup time.

[0108] 3) Based on these 4900 data, calculate the average startup time of each unit τsiavg and the average startup time of all units τsavg:

[0109] Because abnormal data was filtered out in step 2), the number of startups for each unit varies. For example, the startup durations for unit 1# are τs 11, τs 13, …, τs 1100. Because τs 12 was too large and was eliminated, the startup durations for unit 1# total fewer than 100, only 99. Taking the average startup duration for unit 1#, we obtain τs1avg = (τs11 + τs13 + τs14 … + τs1100) / 99. Similarly, we obtain the average startup durations for the other units: τs2avg, τs3avg, …, τs50avg, resulting in 50 values, each representing the average startup duration of a unit.

[0110] Then take the average value of these 4900 τsij to obtain the average startup time τsavg of all units.

[0111] 2.3 Calculation of bearing temperature rise

[0112] In the power generation state, due to the bearing sticking, the friction is too large. Under the same power operation condition, the temperature rise of its bearing is higher than that of other units.

[0113] 1) Identify the power generation state, extract and calculate the transmission chain bearing temperature rise (bearing temperature - ambient temperature) of all units in the site under power generation state, and record the power generation power P corresponding to the temperature rise;

[0114] 2) Divide the temperature rise into bins based on the generated power. Since the power range is 0-Pr (Pr is the rated power, which is also the maximum power), 10% of Pr is used as a bin. Calculate the average bearing temperature rise in the bin. For example, the average bearing temperature rises corresponding to the intervals of 0-10% Pr, 10%-20% Pr, ... 90%-100% Pr are T1i, T2i, ... T10i.

[0115] (i=1,2...N, i represents the unit number), that is, 10 quantities are obtained for each unit;

[0116] 3) Calculate the average bearing temperature rise of all units in each power bin:

[0117] T1avg=(T11+T12+…+T1N) / N;

[0118] T2avg=(T21+T22+…+T2N) / N, where N is the unit number;

[0119] Then the average temperature rise T1avg, T2avg, ...T10avg of all units in each power compartment is obtained;

[0120] 3. Abnormal identification

[0121] The judgment of bearing abnormality requires that three conditions be met at the same time:

[0122] 1) Condition 1: Compare the average standby speed ωiavg of each unit obtained from the above calculation results with the average standby speed ωavg of the entire site. If it exceeds a certain threshold ω1, then condition 1 is considered to be met:

[0123] ωavg-ωiavg=ωDi>ω1

[0124] 2) Condition 2: Compare the average startup time τsiavg of each unit obtained from the above calculation results with the average startup time τsavg of the entire field. If they exceed a certain threshold τs1, condition 2 is considered to be met:

[0125] τsiavg-τsavg=τsDi>τs1

[0126] 3) Condition 3: Compare the average bearing temperature rise of each unit obtained from the above calculation results with the average temperature rise of the entire field. Each sub-bin is required to be compared separately, that is, whether the difference between Tki and Tkavg exceeds T0 (k = 1, 2, ... 10 sub-bins). If 7 sub-bins meet the difference exceeding T0 (here it is assumed that the preset number of sub-bins = 5, that is, 7 out of 10 sub-bin temperatures exceed 5, then it is considered that condition 3 is met.

[0127] 4. Operation and maintenance tips

[0128] 1) Provide operation and maintenance prompts for those who meet conditions 1, 2, and 3 at the same time.

[0129] 2) The threshold T0 in condition 3 can be set to three levels: T1, T2, and T3, with T3>T2>T1:

[0130] When T0 is selected as T1, the bearing abnormality will be reported;

[0131] When T0 is selected as T2, a bearing abnormality warning is issued;

[0132] When T0 is selected as T3, the bearing abnormality is reported urgently;

[0133] On-site operation and maintenance are carried out based on these three levels of information prompts.

[0134] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the steps of the wind turbine bearing sticking anomaly identification method are implemented. The memory may include internal memory, such as high-speed random access memory, and may also include non-volatile memory, such as at least one disk storage device. The processor, network interface, and memory are interconnected via an internal bus. The internal bus may be an industrial standard architecture bus, a peripheral component interconnect standard bus, an extended industrial standard architecture bus, etc. The bus may be divided into an address bus, a data bus, a control bus, etc. The memory is used to store programs. Specifically, the program may include program code, and the program code includes computer operating instructions. The memory may include internal memory and non-volatile memory, and provides instructions and data to the processor.

[0135] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for identifying abnormal bearing seizure in a wind turbine generator set, characterized in that: The following steps are involved: S1. Obtain the unit operation data of the target wind farm and perform data cleaning. The cleaned data includes the standby speed data of each unit, the startup duration data in the startup state, and the bearing temperature rise data; S2. The standby speed data is processed by the standby speed model to output the average standby speed of each unit ωiavg and the average standby speed of all units ωavg; The startup time data in the startup state is passed through the startup time model to output the average startup time of each unit τsiavg and the average startup time of all units τsavg; The bearing temperature rise data is processed by the bearing temperature rise model to output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin; Where i = 1, 2...N, i represents the unit number; k = 1, 2...K, k represents the number of compartments; S3. Determine whether the wind turbine bearing has stuck abnormally based on the following three conditions: Condition 1: Compare the average standby speed of all units, ωavg, with the average standby speed of each unit, ωiavg. If the speed exceeds the threshold, condition 1 is considered to be met. Condition 2: Compare the average startup time of each unit, τsiavg, with the average startup time of all units, τsavg. If the time exceeds the threshold, condition 2 is considered to be met. Condition 3: Compare the average bearing temperature rise (Tki) within each power bin of each unit with the average bearing temperature rise (Tkavg) of all units in each power bin. If the number of bins exceeding the temperature rise threshold exceeds the preset number of bins, condition 3 is considered met. For a certain wind turbine generator set i, when all three conditions are met, it is determined that the bearing of the wind turbine generator set is stuck abnormally.

2. A method for identifying abnormal bearing seizure in a wind turbine generator set according to claim 1, characterized in that: In S2, the standby speed data is processed by the standby speed model to output the average standby speed of each unit ωiavg and the average standby speed of all units ωavg; The specific calculation process of the standby speed model is: 2.11) Identify the status of each unit's operating data and identify the data marked as standby; 2.12) Calculate the average impeller speed ωij of all units in the site during each standby period, where i = 1, 2...N, i represents the unit number; j = 1, 2...M, j represents the number of standby periods; 2.13) Sort the standby speeds ωij of all units on site, remove the x% maximum and x% minimum values, and obtain the filtered data; 2.14) Calculate the average standby speed ωiavg of each unit based on the filtered data, and take the average standby speed ωavg of all units in the field; Then take the average value of the filtered data to obtain the average standby speed ωavg of all units.

3. The method for identifying abnormal bearing seizure in a wind turbine generator set according to claim 1, wherein: In S2, the startup duration data in the startup state is processed by the startup duration model to output the average startup duration τsiavg of each unit and the average startup duration τsavg of all units in the field; the specific calculation process of the startup duration model is: 2.21) Identify the status of each unit's operating data and identify the data marked as startup; Calculate the duration of each startup process τs ij for all units in the site, where i = 1, 2...N, i represents the unit number; j = 1, 2...M, j represents the number of startups; 2.22) Sort the startup durations τs ij of all units on site, remove the x% maximum and x% minimum values, and obtain qualified startup duration data; 2.23) Calculate the qualified startup duration data of each unit to obtain the average startup duration τsiavg of each unit; Then take the average of the qualified startup time data to obtain the average startup time τsavg of all units.

4. The method for identifying bearing sticking anomaly in a wind turbine generator set according to claim 1, characterized in that: In S2, the bearing temperature rise data is processed by the bearing temperature rise model to output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin. The calculation process of the bearing temperature rise model is as follows: 2.31) Identify the power generation state, extract and calculate the temperature rise of the transmission chain bearings of all organic units in the site under the power generation state, and record the power generation power P corresponding to the temperature rise; 2.32) Divide the temperature rise into bins based on the generated power. The power range is 0-Pr, where Pr is the rated power. Divide 0-Pr into k bins and calculate the average bearing temperature rise Tki in each power bin for each unit. 2.33) Calculate the average temperature rise T1avg, T2avg, ..., Tkavg of all units in each power compartment; Tkavg=(Tk1+Tk2+…+TkN) / N, where N is the unit number.

5. The method for identifying bearing sticking anomaly in a wind turbine generator set according to claim 1, characterized in that: In S3, condition 1 is specifically: Compare the average standby speed of all units ωavg with the average standby speed of each unit ωiavg. If they exceed the speed threshold ω1, then condition 1 is considered to be met: ωavg-ωiavg=ωDi>ω1.

6. The method for identifying bearing sticking anomaly in a wind turbine generator set according to claim 1, characterized in that: In S3, condition 2 is specifically: Compare the average startup time of each unit τsiavg with the average startup time of all units τsavg. If it exceeds a certain threshold τs1, it is considered that condition 2 is met: τsiavg-τsavg=τsDi>τs1.

7. The method for identifying bearing sticking anomaly in a wind turbine generator set according to claim 1, characterized in that: In S3, condition 3 is specifically: If the difference between Tki and Tkavg exceeds the temperature rise threshold T0, and the number of compartments exceeding the temperature rise threshold T0 exceeds the preset number of compartments, it is determined that condition 3 is met.

8. The method for identifying bearing sticking anomaly in a wind turbine generator set according to claim 1, characterized in that: Prompt based on the results of S3, specifically: For a certain wind turbine generator set i, if the three conditions are met at the same time, the operation and maintenance prompt of the unit will be issued; The threshold T0 in condition 3 can be set to three levels: T1, T2, and T3, with T3>T2>T1: When T0 is selected as T1, the bearing abnormality will be reported; When T0 is selected as T2, a bearing abnormality warning is issued; When T0 is selected as T3, the bearing abnormality is reported urgently.

9. A wind turbine bearing sticking anomaly identification system, characterized in that: include: The data processing module is used to clean the operating data of the units of the target wind farm. The cleaned data includes the standby speed data of each unit, the startup time data in the startup state, and the bearing temperature rise data; The standby speed model is used to process the standby speed data and output the average standby speed ωiavg of each unit and the average standby speed ωavg of all units in the field; The startup duration model is used to process the startup duration data in the startup state and output the average startup duration τsiavg of each unit and the average startup duration τsavg of all units in the field; The bearing temperature rise model is used to process the bearing temperature rise data and output the average bearing temperature rise Tki in each power bin of each unit and the average bearing temperature rise Tkavg of all units in each power bin; Where i = 1, 2...N, i represents the unit number; k = 1, 2...K, k represents the number of compartments; S3, abnormal identification module, is used to determine whether there is any abnormality in the bearing of the wind turbine generator set: Condition 1: Compare the average standby speed of all units, ωavg, with the average standby speed of each unit, ωiavg. If the speed exceeds the threshold, condition 1 is considered to be met. Condition 2: Compare the average startup time of each unit, τsiavg, with the average startup time of all units, τsavg. If the time exceeds the threshold, condition 2 is considered to be met. Condition 3: Compare the average bearing temperature rise (Tki) within each power bin of each unit with the average bearing temperature rise (Tkavg) of all units in each power bin. If the number of bins exceeding the temperature rise threshold exceeds the preset number of bins, condition 3 is considered met. When all three conditions are met for a certain wind turbine generator set i, it is determined that the bearing of the wind turbine generator set of the set is stuck abnormally.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the method for identifying abnormal bearing sticking of a wind turbine generator set according to any one of claims 1 to 8 are implemented.

Images