Methods and devices for detecting faults in the transmission chain of wind turbine units

Abstract

This application proposes a method and apparatus for detecting faults in the wind turbine drivetrain, relating to the field of fault detection technology. The method includes: acquiring time-series operating data and state waveform data of the wind turbine drivetrain; determining abnormal characteristic indicators of the wind turbine drivetrain based on whether the time-series operating data meets preset target rules; determining time-domain and frequency-domain characteristic indicators of the wind turbine drivetrain based on the state waveform data; determining fused features based on the abnormal characteristic indicators, time-domain characteristic indicators, and frequency-domain characteristic indicators; and determining the fault detection result of the wind turbine drivetrain based on the fused features. This addresses the technical problems of insufficient data feature mining and low fault detection accuracy in existing technologies.

Description

Technical Field

[0001] This application relates to the field of fault detection technology, and in particular to a method and device for detecting faults in the transmission chain of a wind turbine. Background Technology

[0002] The wind turbine drivetrain system is a key component for condition monitoring and fault diagnosis. Fault prediction and health management of the drivetrain are important guarantees for the efficient operation of wind power equipment. Real-time monitoring of the wind turbine drivetrain's operating status and early warning and location of serious faults can help extend the service life of major components in the drivetrain system and reduce power generation losses and operation and maintenance costs.

[0003] Most wind turbines are currently equipped with Supervisory Control and Data Acquisition (SCADA) systems and Condition Monitoring Systems (CMS) to collect various operating parameters of the wind turbine under different modes, such as speed, temperature, and vibration. Most wind turbine drivetrain system fault early warning systems are based on a single signal source, either SCADA or CMS data. However, when a fault occurs in the drivetrain system, there is significant interference and coupling between fault characteristics, and different components in the drivetrain system have obvious correlations. Therefore, when analyzing based on a single signal source, feature mining is not comprehensive enough, and the accuracy of fault detection and early warning is not high. Summary of the Invention

[0004] This application aims to at least partially address one of the technical problems in the related art.

[0005] Therefore, the first objective of this application is to propose a method for detecting faults in the transmission chain of wind turbine units, so as to achieve accurate fault detection.

[0006] The second objective of this application is to propose a fault detection device for wind turbine drivetrain.

[0007] The third objective of this application is to propose an electronic device.

[0008] The fourth objective of this application is to provide a computer-readable storage medium.

[0009] The fifth objective of this application is to provide a computer program product.

[0010] To achieve the above objectives, the first aspect of this application proposes a method for detecting faults in the transmission chain of a wind turbine, comprising:

[0011] Acquire the timing operation data and status waveform data of the wind turbine drive train;

[0012] Based on whether the time-series operation data meets the preset target rules, the abnormal characteristic indicators of the wind turbine transmission chain are determined;

[0013] Based on the state waveform data, determine the time-domain and frequency-domain characteristic indicators of the wind turbine drive train;

[0014] Based on the abnormal feature indicators, the time-domain feature indicators, and the frequency-domain feature indicators, a fusion feature is determined, and the fault detection result of the wind turbine drivetrain is determined according to the fusion feature.

[0015] To achieve the above objectives, a second aspect of this application provides a wind turbine drivetrain fault detection device, comprising:

[0016] The first acquisition module is used to acquire the timing operation data and status waveform data of the wind turbine drive train;

[0017] The second acquisition module is used to determine the abnormal characteristic indicators of the wind turbine transmission chain based on whether the time-series operation data meets the preset target rules.

[0018] The third acquisition module is used to determine the time-domain characteristic index and frequency-domain characteristic index of the wind turbine transmission chain based on the state waveform data.

[0019] The fault detection module is used to determine the fused features based on the abnormal feature indicators, the time-domain feature indicators, and the frequency-domain feature indicators, and to determine the fault detection result of the wind turbine drive train based on the fused features.

[0020] To achieve the above objectives, a third aspect of this application provides an electronic device, including: a processor, and a memory communicatively connected to the processor;

[0021] The memory stores computer-executed instructions;

[0022] The processor executes computer execution instructions stored in the memory to implement the method as described in the first aspect embodiment.

[0023] To achieve the above objectives, a fourth aspect of this application provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, are used to implement the method described in the first aspect.

[0024] To achieve the above objectives, a fifth aspect of this application provides a computer program product including a computer program that, when executed by a processor, implements the method described in the first aspect.

[0025] The wind turbine drivetrain fault detection method and device provided in this application acquires the time-series operation data and state waveform data of the wind turbine drivetrain, matches and judges the abnormality rules in the target rules based on the time-domain operation data, thereby determining the abnormal characteristic indicators of the wind turbine drivetrain, and acquires the time-domain characteristic indicators and frequency-domain characteristic indicators based on the state waveform data, fully mining the features in the state waveform data, and fusing the abnormal characteristic indicators, time-domain characteristic indicators and frequency-domain characteristic indicators to obtain fused features for fault detection, which solves the problem that the current feature mining is not comprehensive enough and the fault detection and early warning accuracy is not high.

[0026] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0027] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0028] Figure 1 A flowchart illustrating a wind turbine drivetrain fault detection method provided in this application embodiment;

[0029] Figure 2 This is a schematic diagram of a process for obtaining abnormal feature indicators provided in an embodiment of this application;

[0030] Figure 3 This is a schematic diagram of a process for obtaining feature indicators and frequency domain feature indicators provided in an embodiment of this application;

[0031] Figure 4 This is a schematic diagram of a process for obtaining fault detection results provided in an embodiment of this application;

[0032] Figure 5 A flowchart illustrating another wind turbine drivetrain fault detection method provided in this application embodiment;

[0033] Figure 6 A logic diagram of a wind turbine drivetrain fault detection method provided in an embodiment of this application;

[0034] Figure 7 This is a schematic diagram of a wind turbine drivetrain fault detection device provided in an embodiment of this application. Detailed Implementation

[0035] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0036] The following describes a wind turbine drivetrain fault detection method and apparatus according to embodiments of this application, with reference to the accompanying drawings.

[0037] Figure 1 This is a flowchart illustrating a wind turbine drivetrain fault detection method provided in an embodiment of this application. Figure 1 As shown, the method includes the following steps:

[0038] S101, acquire the timing operation data and status waveform data of the wind turbine drive train.

[0039] In some implementations, the timing operation data of the wind turbine drive train can be obtained by monitoring the SCADA system, which may include wind turbine operation-related parameters such as temperature, power and speed; optionally, in this embodiment, data can be collected at 1-second intervals to obtain the timing operation data for a day.

[0040] In some implementations, the state waveform data of the wind turbine drivetrain can be obtained by monitoring the CMS system, which may include vibration signal waveform data of various important components such as gearbox, generator and main shaft; optionally, in this embodiment, data can be collected at 2-hour intervals to obtain state waveform data for one day.

[0041] Optionally, the time-series operation data and status waveform data can be data collected from the wind turbine's operation the previous day. The relevant data generated during the wind turbine's operation is stored in a big data platform. After the day's operation ends, the data from the previous day is read from the big data platform in the early morning to obtain the time-series operation data and status waveform data for subsequent analysis. In other embodiments, an analysis cycle can be half a day or longer, that is, time-series operation data and status waveform data within half a day or longer can be obtained for fault detection. This will not be elaborated on in detail here.

[0042] In some implementations, SCADA data and CMS data can be preprocessed separately. For SCADA data, the temperature at the current moment is greatly affected by the previous moment, but less affected by short-term fluctuations in characteristic parameters such as wind speed and power. Therefore, abnormal data with large short-term fluctuations can be removed, and missing data can be supplemented. For CMS data, due to sensor or acquisition channel failure, there may be distortion of the acquired vibration data. In the data preprocessing stage, it is necessary to remove invalid data, such as waveforms with asymmetrical signals, waveforms with very small signal amplitudes, clipping, abnormal impacts, and waveforms with oversaturation impacts.

[0043] S102, determine the abnormal characteristic indicators of the wind turbine drive chain based on whether the time-series operation data meets the preset target rules.

[0044] Optionally, the preset target rules can be rules for determining whether there are abnormal situations, such as abnormal situations like excessively rapid temperature rise in a short period of time, abnormal temperature of the engine drive-end bearing, abnormal temperature of the engine non-drive-end bearing, and abnormal temperature of the gearbox. Based on the temperature data in the time-series operation data, it is determined whether any of the preset target rules are met. If any abnormal rule is met, it indicates that there is an abnormality in the current time-series operation data.

[0045] Optionally, this embodiment can set an exception value of 0 and 1. Based on the time-series running data, each rule in the preset target rules is matched and judged. When the time-series running data meets the corresponding rule, that is, when the exception condition is met, the corresponding exception value is 1; when the time-series running data does not meet the corresponding rule, that is, when it is in the normal condition, the corresponding exception value is 0.

[0046] Understandably, each rule in the preset target rules is matched and judged based on the time-series running data to determine the outlier value corresponding to each rule, thereby obtaining the same number of outliers as the number of rules. All outliers are combined to obtain anomaly feature indicators, for example, using all outliers to form a feature vector, and using the feature vector as anomaly feature indicator.

[0047] S103. Based on the state waveform data, determine the time-domain and frequency-domain characteristic indicators of the wind turbine drive train.

[0048] In some implementations, the state waveform data is the acquired time-domain waveform data. The time-domain waveform data can be converted to obtain the corresponding frequency-domain waveform data, and the frequency-domain characteristic indicators can be calculated from the frequency-domain waveform data.

[0049] Optionally, time-domain characteristic indicators may include indicators that characterize time-domain waveform features, such as effective value, variance, root square amplitude, peak value, and time-domain entropy.

[0050] Optionally, frequency domain characteristic indicators may include frequency domain entropy, envelope spectrum entropy, envelope spectrum peak factor, and meshing sideband energy of different levels, wherein the envelope spectrum entropy and envelope spectrum peak factor can be the upper envelope of the extracted frequency domain waveform data, and the indicators are calculated from the upper envelope.

[0051] In some implementations, wind turbine units may include different key components such as bearings and gearboxes. Bearings include inner and outer ball bearings, while gearboxes include planetary gears and sun gears. Therefore, different frequency domain indicators can be calculated separately from the frequency domain waveforms of the bearings and gearboxes. For example, the frequency domain characteristic indicators of bearings include the peak factor of the inner ring envelope spectrum, the peak factor of the outer ring envelope spectrum, and the peak factor of the rolling element envelope spectrum; the frequency domain characteristic indicators of gearboxes include the low-speed sideband energy of the first-stage meshing, the high-speed sideband energy of the first-stage meshing, and the low-speed sideband energy of the second-stage meshing.

[0052] S104. Based on anomaly characteristic indicators, time-domain characteristic indicators, and frequency-domain characteristic indicators, the fusion characteristics are determined, and the fault detection results of the wind turbine drive chain are determined according to the fusion characteristics.

[0053] Optionally, abnormal feature indicators, time-domain feature indicators, and frequency-domain feature indicators can be concatenated to obtain fused features. In some implementations, basic information and operating parameters of the wind turbine, such as rated speed, wind speed, and generator speed, can also be obtained. The fused features are obtained by fusing the additional feature information with the abnormal feature indicators, time-domain feature indicators, and frequency-domain feature indicators.

[0054] By inputting the fused features into the pre-trained fault identification model, the fault detection results of whether there is a fault in the wind turbine drive chain are obtained, thereby improving the accuracy of fault detection.

[0055] In this embodiment, by acquiring the time-series operation data and state waveform data of the wind turbine drivetrain, and matching and judging the abnormal rules in the target rules based on the time-domain operation data, the abnormal characteristic indicators of the wind turbine drivetrain are determined. Based on the state waveform data, time-domain characteristic indicators and frequency-domain characteristic indicators are acquired, and the features in the state waveform data are fully explored. The abnormal characteristic indicators, time-domain characteristic indicators and frequency-domain characteristic indicators are fused to obtain fused features. Fault detection is performed using more comprehensive and rich fused features to obtain more accurate and reliable fault detection results and improve the accuracy of fault analysis.

[0056] Based on the above embodiments, the process of obtaining abnormal feature indicators is described. In this embodiment, the time-series operation data includes at least one of the following: generator active power, fan speed, generator speed, main shaft speed, unit operating status, generator drive end bearing temperature, generator non-drive end bearing temperature, gearbox oil temperature, inverter power, gearbox drive end bearing temperature and gearbox non-drive end bearing temperature. Figure 2 This is a schematic diagram illustrating a process for obtaining abnormal feature indicators provided in an embodiment of this application. Figure 2 As shown, the method includes:

[0057] S201, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator non-drive end bearing temperature, determine whether the generator non-drive end bearing temperature abnormality rule is met, and obtain the first abnormal value.

[0058] Optionally, the unit's operating status may include grid-connected status, standby status, shutdown status, and maintenance status. This embodiment analyzes the unit's operating status when it is in grid-connected status.

[0059] Optionally, the turbines in the wind farm can be divided into power ranges based on their generator active power to obtain a first power range and a second power range. For example, for a 0.85MW turbine, the first power range based on generator active power is [600, ∞), and the second power range is [0, 600); for a 1.25MW turbine, the first power range based on generator active power is [1000, ∞), and the second power range is [0, 1000). At the current moment, the turbine is determined to be in either the first or second power range based on the range in which its generator active power falls.

[0060] In some implementations, the time interval of the time-series running data can be processed. In this embodiment, the data with a sampling interval of 1 second is averaged and converted into time-series running data with a 1-minute interval. That is, the data within 1 minute is averaged and used as the data within that 1 minute, thereby obtaining time-series running data with a 1-minute interval. In this embodiment, at the same time, the first outlier is obtained at each minute.

[0061] For the first power range, it can be determined that the temperature of the generator non-drive end bearing of the unit is at least higher than the temperature of the generator drive end bearing by a fixed value for a first continuous period of time. In this embodiment, a box plot analysis can be performed on the temperature of the generator non-drive end bearing to determine the temperature of the generator non-drive end bearing is at least higher than the temperature of the generator drive end bearing by a fixed value. For example, if the fixed value is 20 degrees, then the temperature of the generator non-drive end bearing is determined to be at least 20 degrees higher than the temperature of the generator drive end bearing for a first continuous period of time.

[0062] In response to a first continuous duration exceeding a preset time length, the first abnormal value of the unit is determined to be 1; optionally, assuming the preset time length is 10 minutes, if the first continuous duration is greater than 10 minutes, the temperature of the non-drive end bearing of the generator of the unit is determined to be abnormal, and the corresponding first abnormal value is 1; otherwise, if the temperature of the non-drive end bearing of the generator of the unit is not abnormal, the first abnormal value is 0.

[0063] For the second power range, it can be determined whether the temperature of the generator non-drive end bearing of the unit is greater than the second preset temperature; in response to the generator non-drive end bearing temperature being greater than the second preset temperature, the first abnormal value of the unit is determined to be 1; otherwise, the first abnormal value of the unit is determined to be 0; wherein the second preset temperature can be 70 degrees.

[0064] S202, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator drive end bearing temperature, determine whether the generator drive end bearing temperature abnormality rule is met, and obtain the second abnormal value.

[0065] For the first power range, a second continuous period can be determined where the temperature of the generator drive end bearing is at least higher than the temperature of the generator non-drive end bearing by a fixed value. In this embodiment, a box plot analysis can be performed on the temperature of the generator drive end bearing to determine the second continuous period where the temperature of the generator drive end bearing is at least higher than the temperature of the generator non-drive end bearing by a fixed value, for example, if the fixed value is 20 degrees, then the second continuous period where the temperature of the generator drive end bearing is at least 20 degrees higher than the temperature of the generator non-drive end bearing is determined.

[0066] In response to the second continuous duration being longer than a preset time length, the first abnormal value of the unit is determined to be 1; optionally, if the preset time length is 10 minutes, then if the second continuous duration is longer than 10 minutes, the temperature of the generator drive end bearing of the unit is determined to be abnormal, and the corresponding first abnormal value is 1; otherwise, if the temperature of the generator drive end bearing of the unit is not abnormal, the first abnormal value is 0.

[0067] For the second power range, it can be determined whether the temperature of the generator drive end bearing of the unit is greater than the second preset temperature; in response to the generator drive end bearing temperature of the unit being greater than the second preset temperature, the first abnormal value of the unit is determined to be 1; otherwise, the first abnormal value of the unit is determined to be 0; wherein the second preset temperature can be 70 degrees.

[0068] S203, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and gearbox oil temperature, determines whether the gearbox oil temperature abnormality rule is met, and obtains the third abnormal value.

[0069] For the first power range, the third consecutive duration for which the gearbox oil temperature of the unit is higher than the first preset temperature can be determined. In this embodiment, a box plot analysis can be performed on the gearbox oil temperature to determine the third consecutive duration for which the gearbox oil temperature is higher than the first preset temperature. For example, the first preset temperature is the sum of the average gearbox temperature and a fixed value, where the fixed value is 20 degrees. Then, the third consecutive duration for which the gearbox temperature is higher than the average temperature value + 20 degrees is determined.

[0070] In response to the third continuous duration being longer than a preset time length, the first abnormal value of the unit is determined to be 1; optionally, if the preset time length is 10 minutes, then if the third continuous duration is longer than 10 minutes, the gearbox oil temperature of the unit is determined to be abnormal, and the corresponding first abnormal value is 1; otherwise, if the gearbox oil temperature of the unit is not abnormal, the first abnormal value is 0.

[0071] For the second power range, it can be determined whether the gearbox oil temperature of the unit is greater than the second preset temperature; in response to the gearbox oil temperature of the unit being greater than the second preset temperature, the first abnormal value of the unit is determined to be 1; otherwise, the first abnormal value of the unit is determined to be 0; wherein the second preset temperature can be 70 degrees.

[0072] S204. Based on the inverter power and the gearbox drive end bearing temperature, determine whether the gearbox drive end bearing temperature abnormality rule is met, and obtain the fourth abnormal value.

[0073] The inverter power collected per second is obtained, and the inverter power of the unit is determined based on the inverter power to see if it is greater than the preset power. Optionally, the preset power can be 2000kW, that is, to determine whether the inverter power is greater than 2000kW.

[0074] In response to the inverter power being greater than a preset power, a fourth continuous duration for which the gearbox drive end bearing temperature is greater than a third preset temperature is determined; in this embodiment, the third preset temperature can be 67.5 degrees, that is, when the inverter power is greater than 2000kW, a fourth continuous duration for which the gearbox drive end bearing temperature is greater than 67.5 degrees is determined.

[0075] In response to the fourth continuous duration being longer than a preset time length, the fourth abnormal value of the unit is determined to be 1; in this embodiment, the preset time length is 10 minutes, that is, when the fourth continuous duration is longer than 10 minutes, the temperature of the gearbox drive end bearing of the unit is determined to be abnormal, and the fourth abnormal value is 1; otherwise, the temperature of the gearbox drive end bearing of the unit is determined to be normal, and the fourth abnormal value is 0.

[0076] S205, based on the inverter power and the temperature of the non-drive end bearing of the gearbox, determine whether the abnormal temperature rule of the non-drive end bearing of the gearbox is met, and obtain the fifth abnormal value.

[0077] The inverter power collected per second is obtained, and the inverter power of the unit is determined based on the inverter power collected per second. Optionally, the preset power can be 2000kW, that is, to determine whether the inverter power is greater than 2000kW.

[0078] In response to the inverter power being greater than a preset power, a fifth consecutive duration is determined for the gearbox non-drive end bearing temperature to be greater than a third preset temperature. In this embodiment, the third preset temperature can be 67.5 degrees, which means that when the inverter power is greater than 2000kW, the fifth consecutive duration for the gearbox non-drive end bearing temperature to be greater than 67.5 degrees is determined.

[0079] In response to the fifth continuous duration being greater than a preset time length, the fifth abnormal value of the unit is determined to be 1; in this embodiment, the preset time length is 10 minutes, that is, when the fifth continuous duration is greater than 10 minutes, the temperature of the non-drive end bearing of the gearbox of the unit is determined to be abnormal, and the fifth abnormal value is 1; otherwise, it is determined that the temperature of the non-drive end bearing of the gearbox of the unit is not abnormal, and the fifth abnormal value is 0.

[0080] S206, obtain the wind turbine startup time, and determine whether the abnormal rule of excessively rapid temperature rise of the generator drive end bearing is met based on the startup time and the degree of temperature change of the generator drive end bearing. The sixth abnormal value is obtained.

[0081] Optionally, the generator set can be divided into an initial startup phase and a normal operation phase based on the startup duration. In this embodiment, the time period from startup to one hour of grid-connected operation is defined as the initial startup phase; and the power generation state after one hour of grid-connected operation is defined as the normal operation phase.

[0082] For units in the initial startup phase, the first change value of the generator drive end bearing temperature within a fixed time period is determined. If the first change value is greater than a first preset change value for two consecutive fixed time periods, the sixth abnormality value of the unit is determined to be 1. Assuming the fixed time period is 30 seconds, the temperature rise of the generator drive end bearing within 30 seconds is taken as the first change value. If the first change value within 30 seconds is greater than the first preset change value twice consecutively, an abnormality is determined for the unit. For example, if the first preset change value is 0.5 degrees Celsius, then if the first change value within 30 seconds is greater than 0.5 degrees Celsius twice consecutively, it is determined that the unit has an abnormal situation of excessively rapid temperature rise in the generator drive end bearing, and the sixth abnormality value is 1. Conversely, if the first preset change value is less than 0.5 degrees Celsius, it is determined that the unit does not have an abnormal situation of excessively rapid temperature rise in the generator drive end bearing, and the sixth abnormality value is 0.

[0083] For a generator unit in normal operation, the third change value of the generator drive end bearing temperature is determined within a fixed time period. If the third change value is greater than the second preset change value for two consecutive fixed time periods, the sixth abnormal value of the generator unit is determined to be 1. Optionally, the fixed time period can be 30 seconds, and the second preset change value is 0.3 degrees. If the temperature rise value (third change value) is greater than 0.3 degrees for two consecutive 30-second periods, it is determined that the generator drive end bearing temperature rises too quickly, and the sixth abnormal value is 1. Otherwise, it is determined that the generator drive end bearing temperature rises too quickly, and the sixth abnormal value is 0.

[0084] S207. Based on the start-up time and the degree of temperature change of the generator non-drive end bearing, determine whether the abnormal rule of excessively rapid temperature rise of the generator non-drive end bearing is met, and obtain the seventh abnormal value.

[0085] For the unit in the initial startup phase, a second change value of the generator non-drive end bearing temperature is determined within a fixed time period. In response to the second change value being greater than a first preset change value for two consecutive fixed time periods, the seventh abnormal value of the unit is determined to be 1. Optionally, the fixed time period can be 30 seconds, and the first preset change value is 0.5 degrees. Then, if the temperature rise value (second change value) within two consecutive 30-second periods is greater than 0.5 degrees, it is determined that the unit has an abnormal situation of excessively rapid temperature rise of the generator non-drive end bearing, and the seventh abnormal value is 1. Conversely, it is determined that the unit does not have an abnormal situation of excessively rapid temperature rise of the generator non-drive end bearing, and the seventh abnormal value is 0.

[0086] For a generator unit in normal operation, the fourth change value of the generator non-drive end bearing temperature is determined within a fixed time period. If the fourth change value is greater than the second preset change value for two consecutive fixed time periods, the seventh abnormal value of the unit is determined to be 1. Optionally, the fixed time period can be 30 seconds, and the second preset change value is 0.3 degrees. If the temperature rise value (fourth change value) is greater than 0.3 degrees for two consecutive 30-second periods, it is determined that the generator unit has an abnormal situation of excessively rapid temperature rise of the generator non-drive end bearing, and the seventh abnormal value is 1. Otherwise, it is determined that the generator unit does not have an abnormal situation of excessively rapid temperature rise of the generator non-drive end bearing, and the seventh abnormal value is 0.

[0087] S208, Based on the first to the seventh outlier, determine the abnormal characteristic indicators of the wind turbine drive train.

[0088] It is understood that the execution order of steps S201-S207 in this embodiment can be changed or executed synchronously. This embodiment is only an example and the specific execution order is not limited. Each abnormal value is 0 or 1. The abnormal characteristic index of the wind turbine drive chain is determined based on the results of the seven abnormal values ​​from the first abnormal value to the seventh abnormal value.

[0089] Based on the methods for obtaining the first to seventh outliers described above, a table corresponding to the target rules can be constructed. This table includes the anomaly judgment rules for generator non-drive end bearing temperature anomaly, generator drive end bearing temperature anomaly, gearbox oil temperature anomaly, gearbox drive end bearing temperature anomaly, gearbox non-drive end bearing temperature anomaly, generator drive end bearing temperature rising too quickly, and generator non-drive end bearing temperature rising too quickly, as detailed in the table below:

[0090] Table 1

[0091]

[0092] Furthermore, after identifying the outlier characteristics, from the first to the seventh outlier, these outlier characteristics can be presented in a tabular format, as shown in the table below:

[0093] Table 2

[0094]

[0095] In this embodiment, the generator active power, unit operating status, generator drive end bearing temperature, generator non-drive end bearing temperature, gearbox oil temperature, inverter power, gearbox drive end bearing temperature, and gearbox non-drive end bearing temperature are judged based on the time-series operation data to determine whether the corresponding judgment rules for abnormal situations are met. The abnormal situation is judged according to the pre-set target rules, which improves the efficiency and accuracy of abnormality identification. Furthermore, abnormal feature indicators are obtained according to various abnormal situations to fully explore the abnormal features in the time-series operation data and improve the accuracy of subsequent analysis.

[0096] Based on the above embodiments, the process of acquiring time-domain characteristic indicators and frequency-domain characteristic indicators is described. The state waveform data includes at least one of the following: radial acceleration time-domain waveform of the main shaft impeller side bearing, radial acceleration time-domain waveform of the main shaft motor side bearing, radial acceleration time-domain waveform of the gearbox first-stage internal gear ring, radial acceleration time-domain waveform of the gearbox low-speed shaft, radial acceleration time-domain waveform of the gearbox high-speed shaft, radial acceleration time-domain waveform of the generator drive end, and radial acceleration time-domain waveform of the generator non-drive end. Figure 3 This is a schematic diagram illustrating a process for obtaining feature indicators and frequency domain feature indicators, provided as an embodiment of this application. Figure 3 As shown, the method includes:

[0097] S301, for any time-domain waveform in the state waveform data, obtain the effective value, variance, root square magnitude, skewness index, kurtosis index, kurtosis factor, waveform index, peak index, margin index, peak value, peak-to-peak value, and time-domain entropy of the time-domain waveform as time-domain feature indicators.

[0098] It is understandable that the state waveform data is data collected within a day at 2-hour intervals. Based on the data from each moment of the day, corresponding time-domain waveforms are generated. These time-domain waveforms can be represented as follows: In some implementations, time-domain waveforms can be encrypted and compressed and stored in a big data platform. Therefore, when reading time-domain waveforms from a big data platform for analysis, preprocessing such as decompression and decryption is performed on the time-domain waveforms to ensure that the time-domain waveforms are visualized and editable.

[0099] For example, the calculation of mean, variance, kurtosis, and skewness can be expressed as follows:

[0100]

[0101]

[0102]

[0103]

[0104] in, The mean; For variance; For raucousness; y is the skewness; N is the total number of data points in the time-domain waveform; The first time-domain waveform i Data at any given moment.

[0105] Furthermore, based on the time-domain waveform, indices such as RMS value, root square amplitude, kurtosis factor, waveform index, peak index, peak value, peak-to-peak value, and time-domain entropy are calculated. The RMS value reflects the average power or energy of the waveform within one cycle; the root square amplitude is the arithmetic square root of the average of the sum of squares of data in the time-domain waveform; the kurtosis factor is an index representing the smoothness of the waveform, used to describe the distribution of variables; the waveform index can be the ratio of the RMS value to the overall average value; the peak index usually refers to the maximum value of the signal within a time interval; the peak value refers to the maximum value of a single peak in the time-domain waveform; the peak-to-peak value refers to the difference between the maximum and minimum values ​​of the signal within a time interval, i.e., the range between the highest and lowest values ​​of the signal; time-domain entropy is a parameter describing the complexity of the signal; it represents the randomness and uncertainty of the signal in the time domain. The calculation and acquisition of the above indices can be achieved according to existing technical solutions, and no limitations are imposed here, nor will they be elaborated further.

[0106] S302 converts the time-domain waveform into a frequency-domain waveform, and determines the frequency-domain entropy, upper envelope, and corresponding unit components of the frequency-domain waveform. The envelope spectrum entropy is determined based on the upper envelope.

[0107] Optionally, a Fourier transform can be used to convert the time-domain waveform into a frequency-domain waveform. In some implementations, window features, including rectangular windows, Blackman windows, Hanning windows, triangular windows, Gaussian windows, and Hamming windows, can be calculated before the Fourier transform to prevent frequency leakage during the Fourier transform.

[0108] In some implementations, the frequency domain entropy is determined based on the frequency domain waveform. Frequency domain entropy is used as an indicator to measure the degree of disorder in the frequency domain signal. It can be calculated from the Fast Fourier Transform (FFT) coefficients of the frequency domain waveform. The FFT coefficients can be obtained from the frequency domain waveform through the Fast Fourier Transform.

[0109] In some implementations, the frequency domain waveform can be mean-removed and normalized, and the peak values ​​of each waveform period can be extracted and connected to obtain the upper envelope of the frequency domain waveform. The envelope spectral entropy is then calculated based on this upper envelope. Envelope spectral entropy is a feature extraction method used to evaluate signal complexity, mainly used to analyze the complexity and trend of signal changes. The specific calculation method will not be elaborated here.

[0110] Optionally, the unit components include bearings and gearboxes. The bearings include an inner ring, an outer ring, and balls, and the gearboxes include planetary gears and a sun gear. Therefore, when the frequency domain waveform belongs to different unit components, the calculated frequency domain characteristic indicators are different. It is understood that all frequency domain characteristic indicators in this embodiment can be calculated based on existing technologies, and this is not limited here, nor will it be described in detail.

[0111] S303, responding to the frequency domain waveform belonging to the bearing, obtains the peak factor of the inner ring envelope spectrum, the peak factor of the outer ring envelope spectrum, the peak factor of the rolling element envelope spectrum, and the peak factor of the cage envelope spectrum based on the upper envelope, and combines the envelope spectrum entropy and the frequency domain entropy to form the frequency domain characteristic index.

[0112] Based on the calculation of the time-domain and frequency-domain characteristic indices of the bearing, all the time-domain and frequency-domain characteristic indices corresponding to the bearing are obtained, as shown in the table below:

[0113] Table 3 Time-domain and frequency-domain characteristic indicators of bearings

[0114]

[0115] S304, responding to the frequency domain waveform belonging to the gearbox, acquires the first-level meshing low-speed sideband energy, first-level meshing high-speed sideband energy, second-level meshing low-speed sideband energy, second-level meshing high-speed sideband energy, third-level meshing low-speed sideband energy, and third-level meshing high-speed sideband energy of the frequency domain waveform, and combines them with the envelope spectral entropy and frequency domain entropy to form frequency domain characteristic indicators.

[0116] Based on the calculation of the time-domain and frequency-domain characteristic indices of the gearbox, all the time-domain and frequency-domain characteristic indices corresponding to the gearbox are obtained, as shown in the table below:

[0117] Table 4. Time-domain and frequency-domain characteristic indicators of the gearbox

[0118]

[0119] In this embodiment, comprehensive time-domain feature indicators are extracted from each time-domain waveform in the state waveform data. The time-domain waveform is then transformed into a frequency-domain waveform to obtain the corresponding frequency-domain waveform. The features in the frequency-domain waveform are mined to obtain the frequency-domain indicator features corresponding to the gearbox and bearing, thereby extracting complete time-domain and frequency-domain feature indicators from the CMS data and improving the accuracy of subsequent fault detection.

[0120] Based on the above embodiments, the process of obtaining fault detection results will be described. Figure 4 This is a schematic diagram illustrating a process for obtaining fault detection results, provided in an embodiment of this application. Figure 4 As shown, the method includes:

[0121] S401, the abnormal feature indicators, time domain feature indicators and frequency domain feature indicators are concatenated to obtain the fused features.

[0122] S402, the fused features are input into the pre-trained fault detection model to obtain the fault detection results output by the fault detection model.

[0123] In some implementations, basic information and operating parameters of the generator set can also be obtained. The basic information includes one or more of the following: power plant data, wind turbine, measuring point, rated speed, global encoding of vibration waveform, number of sampling points, and time. The operating parameters include one or more of the following: generator speed, wind speed, active power, and main shaft speed.

[0124] In some implementations, to better adapt to the actual operating status of the unit and real-time data acquisition, external input parameters can be determined before analyzing the wind turbine, such as power plant-related data, rated speed, and gearbox speed ratio. The failure frequencies of major components in the wind turbine drivetrain can also be obtained in advance and stored in a database. These failure frequencies may include, but are not limited to: inner ring failure frequency, outer ring failure frequency, rolling element failure frequency, cage failure frequency, input shaft, first-stage planetary gear, first-stage sun gear shaft, second-stage planetary gear, second-stage sun gear shaft, high-speed shaft, first-stage planetary gear meshing, second-stage planetary gear meshing, and high-speed gear meshing. Furthermore, the corresponding main shaft parameters of the unit can be obtained, including outer ring diameter, inner ring diameter, rolling element diameter, pitch circle diameter, and number of rolling elements. Gearbox-related parameters, such as first-stage, second-stage, and third-stage gearbox parameters, can also be obtained to calculate relevant rotational frequencies, meshing frequencies, and failure frequencies, and store them in a database.

[0125] Optionally, the rotational frequency can be calculated based on the rotational speed, which is equal to the rotational speed divided by 60; the waveform duration can be calculated based on the number of sampling points and the sampling frequency, which is equal to the number of sampling points divided by the sampling frequency, thereby matching data such as wind speed and generator power at the corresponding time point. The operating parameters such as wind speed and generator power can be the values ​​of the most recent moment within a 5-minute range of the corresponding waveform duration.

[0126] Furthermore, basic information, operating parameters, and fused features can be input into a pre-trained fault detection model to obtain fault detection results for the wind turbine drivetrain. This involves extracting features from various data collected from the previous day's wind turbine operation to obtain fused features, and then analyzing these fused features to obtain fault detection results, thereby improving the accuracy of fault detection for wind turbines. The fault detection results include: main bearing damage, generator bearing damage, gearbox tooth surface damage, gearbox parallel stage bearing damage, misalignment, and absence of faults.

[0127] Optionally, the fault detection model can adopt the Extreme Gradient Boosting (XGBoost) model. Compared with other learning algorithms in the prior art, the advantage of XGBoost is that it performs a second-order Taylor expansion of the loss function, combines the prediction term and the regularization term, and optimizes computational resources by simplifying the objective function to balance the decrease of the loss function and the complexity of the model, and obtains the optimal solution. At the same time, XGBoost automatically performs parallel computation, ensuring the optimal computation speed.

[0128] Understandably, during the training phase of the fault prediction model, a relatively balanced dataset of normal samples and fault samples can be constructed. For example, a dataset can be constructed using data from the past year. The model is trained based on both the normal and fault sample datasets, and fault types are labeled as 0-5 (6 types in total), corresponding to main bearing damage, generator bearing damage, gearbox tooth surface damage, gearbox parallel stage bearing damage, misalignment, and normal, respectively. 70% of the samples in the dataset are used as the training set, 10% as the validation set, and 20% as the test set. These samples are then input into the XGBoost model for training to obtain the fault prediction model. To select a fault prediction model with better performance, this embodiment uses Grid Search Cross-Validation (Grid Search Cross-Validation). The CV method is used to fine-tune the main parameters of the model. By extracting abnormal feature indicators, time-domain feature indicators, frequency-domain feature indicators, as well as basic information and operating parameters from the sample data, a fusion feature is constructed and then input into the fault prediction model for classification training. This enables the full and efficient use of information from different sources. The training effect of the wind turbine drive chain system state classification model can also be evaluated by three evaluation indicators: accuracy, recall, and F1-score, to ensure the fault identification effect.

[0129] In this embodiment, the abnormal feature indicators extracted from SCADA data and the time-domain and frequency-domain feature indicators extracted from CMS data are concatenated to obtain fused features. This also allows for the acquisition of basic information and operating parameters of the unit, better adapting to the actual operating status and real-time data of the unit. The fused features, along with the basic information and operating parameters, are input into a pre-trained fault identification model, which then outputs specific fault identification results, such as damage to the main bearing, generator bearing, gearbox tooth surface, gearbox parallel stage bearing, misalignment, or no fault, thereby improving the efficiency and accuracy of fault identification.

[0130] Based on the above embodiments, Figure 5 This is a flowchart illustrating another wind turbine drivetrain fault detection method provided in an embodiment of this application; as shown below. Figure 5 As shown, the method includes the following steps:

[0131] S501 acquires the timing operation data and status waveform data of the wind turbine drive train.

[0132] S502, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator non-drive end bearing temperature, determines whether the generator non-drive end bearing temperature abnormality rule is met, and obtains the first abnormal value.

[0133] S503, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator drive end bearing temperature, determines whether the generator drive end bearing temperature abnormality rule is met, and obtains the second abnormal value.

[0134] S504, in response to the unit operating status being grid-connected, for the same moment, based on the generator active power and gearbox oil temperature, determines whether the gearbox oil temperature abnormality rule is met, and obtains the third abnormal value.

[0135] S505, based on the inverter power and the gearbox drive end bearing temperature, determines whether the gearbox drive end bearing temperature abnormality rule is met, and obtains the fourth abnormal value.

[0136] S506, based on the inverter power and the temperature of the non-drive end bearing of the gearbox, determines whether the abnormal temperature rule of the non-drive end bearing of the gearbox is met, and obtains the fifth abnormal value.

[0137] S507: Obtain the wind turbine startup time. Based on the startup time and the degree of temperature change of the generator drive end bearing, determine whether the abnormal rule of excessively rapid temperature rise of the generator drive end bearing is met, and obtain the sixth abnormal value.

[0138] S508, based on the start-up time and the degree of temperature change of the generator non-drive end bearing, determines whether the abnormal rule of excessively rapid temperature rise of the generator non-drive end bearing is met, and obtains the seventh abnormal value.

[0139] S509, Based on the first to the seventh outlier, determine the abnormal characteristic indicators of the wind turbine drive train.

[0140] S510: For any time-domain waveform in the state waveform data, obtain the effective value, variance, root square magnitude, skewness index, kurtosis index, kurtosis factor, waveform index, peak index, margin index, peak value, peak-to-peak value, and time-domain entropy of the time-domain waveform as time-domain feature indicators.

[0141] S511 converts the time-domain waveform into a frequency-domain waveform, and determines the frequency-domain entropy, upper envelope, and corresponding unit components of the frequency-domain waveform. The envelope spectrum entropy is determined based on the upper envelope.

[0142] S512, in response to the frequency domain waveform belonging to the bearing, obtains the peak factor of the inner ring envelope spectrum, the peak factor of the outer ring envelope spectrum, the peak factor of the rolling element envelope spectrum, and the peak factor of the cage envelope spectrum based on the upper envelope, and combines the envelope spectrum entropy and the frequency domain entropy to form the frequency domain characteristic index.

[0143] S513, responding to the frequency domain waveform belonging to the gearbox, acquires the first-level meshing low-speed sideband energy, first-level meshing high-speed sideband energy, second-level meshing low-speed sideband energy, second-level meshing high-speed sideband energy, third-level meshing low-speed sideband energy, and third-level meshing high-speed sideband energy of the frequency domain waveform, and combines them with the envelope spectral entropy and frequency domain entropy to form frequency domain characteristic indicators.

[0144] S514 concatenates the anomaly feature indicators, time-domain feature indicators, and frequency-domain feature indicators to obtain the fused features.

[0145] S515 inputs the fused features into the pre-trained fault detection model to obtain the fault detection results output by the fault detection model.

[0146] In this application embodiment, the implementation methods of steps S501-S515 can be implemented in any of the embodiments of this disclosure, and no limitation is made here, nor will it be described in detail.

[0147] In this embodiment, by acquiring the time-series operation data and state waveform data of the wind turbine drivetrain, the system determines whether the corresponding abnormal situation judgment rules are met based on the time-series operation data. Abnormal situations are judged according to pre-set target rules, improving the efficiency and accuracy of abnormal identification and fully mining the abnormal features in the time-series operation data. Comprehensive time-domain feature indicators are extracted from each time-domain waveform in the state waveform data. The time-domain waveforms are further transformed into frequency-domain waveforms to obtain the corresponding frequency-domain waveforms. Features in the frequency-domain waveforms are mined to obtain the frequency-domain indicator features corresponding to the gearbox and bearings. The abnormal feature indicators extracted from SCADA data and the time-domain and frequency-domain feature indicators extracted from CMS data are concatenated to obtain fused features. Basic information and operating parameters of the unit can also be obtained, better adapting to the actual operating status and real-time data of the unit. The fused features, along with the basic information and operating parameters, are input into a pre-trained fault identification model, which outputs specific fault identification results, improving the efficiency and accuracy of fault identification.

[0148] Figure 6 This is a logical schematic diagram of a wind turbine drivetrain fault detection method provided in this application embodiment. SCADA data and CMS data are acquired separately, and preprocessing operations such as data cleaning are performed on the SCADA and CMS data to improve data analysis accuracy. Anomaly judgment is performed based on the SCADA data and preset target rules to obtain anomaly feature indicators. Time-domain and frequency-domain feature extraction is performed based on the CMS data. Finally, the anomaly feature indicators, time-domain feature indicators, and frequency-domain feature indicators are fused to obtain fused features. These fused features are then input into a pre-trained fault detection model to output fault detection results. This method fully mines the relevant data features of the wind turbine, improving the accuracy of fault identification.

[0149] To achieve the above embodiments, this application also proposes a wind turbine drivetrain fault detection device.

[0150] Figure 7 This is a schematic diagram of a wind turbine drivetrain fault detection device provided in an embodiment of this application. Figure 7 As shown, the wind turbine drivetrain fault detection device 700 includes:

[0151] The first acquisition module 701 is used to acquire the timing operation data and status waveform data of the wind turbine drive train;

[0152] The second acquisition module 702 is used to determine the abnormal characteristic indicators of the wind turbine drive chain based on whether the time-series operation data meets the preset target rules.

[0153] The third acquisition module 703 is used to determine the time-domain and frequency-domain characteristic indicators of the wind turbine drive train based on the state waveform data.

[0154] The fault detection module 704 is used to determine the fused features based on abnormal feature indicators, time domain feature indicators and frequency domain feature indicators, and to determine the fault detection results of the wind turbine drive chain based on the fused features.

[0155] Furthermore, in one possible implementation of this application embodiment, the timing operation data includes at least one of the following: generator active power, fan speed, generator speed, main shaft speed, unit operating status, generator drive end bearing temperature, generator non-drive end bearing temperature, gearbox oil temperature, inverter power, gearbox drive end bearing temperature, and gearbox non-drive end bearing temperature.

[0156] Furthermore, in one possible implementation of this application embodiment, the second acquisition module 702 includes:

[0157] In response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator non-drive end bearing temperature, it is determined whether the generator non-drive end bearing temperature abnormality rule is met, and the first abnormal value is obtained.

[0158] In response to the unit operating status being grid-connected, for the same moment, based on the generator active power and the generator drive end bearing temperature, it is determined whether the generator drive end bearing temperature abnormality rule is met, and a second abnormal value is obtained.

[0159] In response to the unit operating status being grid-connected, for the same moment, based on the generator active power and gearbox oil temperature, it is determined whether the gearbox oil temperature abnormality rule is met, and the third abnormal value is obtained;

[0160] Based on the inverter power and the gearbox drive end bearing temperature, determine whether the gearbox drive end bearing temperature anomaly rule is met, and obtain the fourth anomaly value;

[0161] Based on the inverter power and the temperature of the non-drive end bearing of the gearbox, determine whether the abnormal temperature rule of the non-drive end bearing of the gearbox is met, and obtain the fifth abnormal value;

[0162] The wind turbine startup time is obtained. Based on the startup time and the degree of temperature change of the generator drive end bearing, it is determined whether the abnormal rule of excessively rapid temperature rise of the generator drive end bearing is met, and the sixth abnormal value is obtained.

[0163] Based on the start-up time and the degree of temperature change of the non-drive end bearing of the generator, it is determined whether the abnormal rule of excessively rapid temperature rise of the non-drive end bearing of the generator is met, and the seventh abnormal value is obtained.

[0164] Based on the first to seventh outliers, abnormal characteristic indicators of the wind turbine drive train were determined.

[0165] Furthermore, in one possible implementation of this application embodiment, the second acquisition module 702 includes:

[0166] The wind farm's turbines are divided into power ranges based on the generator's active power, resulting in a first power range and a second power range.

[0167] For the first power range, a first continuous duration is determined where the temperature of the generator non-drive end bearing of the unit is at least higher than a fixed value of the temperature of the generator drive end bearing, a second continuous duration where the temperature of the generator drive end bearing of the unit is at least higher than a fixed value of the temperature of the generator non-drive end bearing, and a third continuous duration where the temperature of the gearbox oil of the unit is higher than a first preset temperature.

[0168] In response to a first continuous duration exceeding a preset time length, the first abnormal value of the unit is determined to be 1; in response to a second continuous duration exceeding a preset time length, the second abnormal value of the unit is determined to be 1; in response to a third continuous duration exceeding a preset time length, the third abnormal value of the unit is determined to be 1.

[0169] For the second power range, in response to the generator non-drive end bearing temperature being greater than the second preset temperature, the first abnormal value of the unit is determined to be 1; in response to the generator drive end bearing temperature being greater than the second preset temperature, the second abnormal value of the unit is determined to be 1; in response to the gearbox oil temperature being greater than the second preset temperature, the third abnormal value of the unit is determined to be 1.

[0170] Based on the inverter power, determine whether the inverter power of the unit is greater than the preset power. In response to the inverter power being greater than the preset power, determine the fourth continuous duration for the gearbox drive end bearing temperature being greater than the third preset temperature and the fifth continuous duration for the gearbox non-drive end bearing temperature being greater than the third preset temperature.

[0171] In response to the fourth continuous duration being longer than the preset time length, the fourth abnormal value of the unit is determined to be 1; in response to the fifth continuous duration being longer than the preset time length, the fifth abnormal value of the unit is determined to be 1.

[0172] The unit is divided into an initial startup phase and a normal operation phase based on the startup duration.

[0173] For the unit in the initial stage of startup, a first change value of the generator drive end bearing temperature and a second change value of the generator non-drive end bearing temperature within a fixed time period are determined; in response to the first change value being greater than a first preset change value for two consecutive fixed time periods, a sixth abnormal value of the unit is determined to be 1; in response to the second change value being greater than the first preset change value for two consecutive fixed time periods, a seventh abnormal value of the unit is determined to be 1.

[0174] For a generator unit in normal operation, a third change value of the generator drive end bearing temperature and a fourth change value of the generator non-drive end bearing temperature within a fixed time period are determined. In response to the third change value being greater than a second preset change value for two consecutive fixed time periods, a sixth abnormal value of 1 is determined for the generator unit. In response to the fourth change value being greater than a second preset change value for two consecutive fixed time periods, a seventh abnormal value of 1 is determined for the generator unit.

[0175] Furthermore, in one possible implementation of this application embodiment, the state waveform data includes at least one of the following: radial acceleration time-domain waveform of the main shaft impeller side bearing, radial acceleration time-domain waveform of the main shaft motor side bearing, radial acceleration time-domain waveform of the gearbox first-stage internal gear ring, radial acceleration time-domain waveform of the gearbox low-speed shaft, radial acceleration time-domain waveform of the gearbox high-speed shaft, radial acceleration time-domain waveform of the generator drive end, and radial acceleration time-domain waveform of the generator non-drive end.

[0176] Furthermore, in one possible implementation of this application embodiment, the third acquisition module 703 includes:

[0177] For any time-domain waveform in the state waveform data, obtain the effective value, variance, root square magnitude, skewness index, kurtosis index, kurtosis factor, waveform index, peak index, margin index, peak value, peak-to-peak value, and time-domain entropy of the time-domain waveform as time-domain feature indicators.

[0178] The time-domain waveform is converted into a frequency-domain waveform, and the frequency-domain entropy, upper envelope, and corresponding unit components of the frequency-domain waveform are determined. The unit components include bearings and gearboxes.

[0179] Determine the envelope spectrum entropy based on the upper envelope line;

[0180] Since the frequency domain waveform belongs to the bearing, the peak factor of the inner ring envelope spectrum, the peak factor of the outer ring envelope spectrum, the peak factor of the rolling element envelope spectrum, and the peak factor of the cage envelope spectrum are obtained based on the upper envelope. The peak factor of the envelope spectrum and the frequency domain entropy are combined to form the frequency domain characteristic index.

[0181] Since the frequency domain waveform belongs to the gearbox, the energy of the first-level meshing low-speed sideband, the energy of the first-level meshing high-speed sideband, the energy of the second-level meshing low-speed sideband, the energy of the second-level meshing high-speed sideband, the energy of the third-level meshing low-speed sideband, and the energy of the third-level meshing high-speed sideband are obtained, and combined with the envelope spectral entropy and the frequency domain entropy to form the frequency domain characteristic index.

[0182] Furthermore, in one possible implementation of this application embodiment, the fault detection module 704 includes:

[0183] The abnormal feature index, time domain feature index and frequency domain feature index are concatenated to obtain the fused feature.

[0184] Furthermore, in one possible implementation of this application embodiment, the fault detection module 704 includes:

[0185] The fused features are input into the pre-trained fault detection model to obtain the fault detection results output by the fault detection model. The fault detection results include: main bearing damage, generator bearing damage, gearbox tooth surface damage, gearbox parallel stage bearing damage, misalignment, and no fault.

[0186] Furthermore, in one possible implementation of this application embodiment, the device 700 further includes:

[0187] Acquire basic information and operating parameters of the generator set. Basic information includes one or more of the following: power plant, wind turbine, measuring point, rated speed, global encoding of vibration waveform, number of sampling points, and time. Operating parameters include one or more of the following: generator speed, wind speed, active power, and main shaft speed.

[0188] By inputting basic information, operating parameters, and fused features into a pre-trained fault detection model, fault detection results of the wind turbine drivetrain are obtained.

[0189] It should be noted that the foregoing explanation of the embodiment of the wind turbine drivetrain fault detection method also applies to the wind turbine drivetrain fault detection device of this embodiment, and will not be repeated here.

[0190] In this embodiment, by acquiring the time-series operation data and state waveform data of the wind turbine drivetrain, the system determines whether the corresponding abnormal situation judgment rules are met based on the time-series operation data. Abnormal situations are judged according to pre-set target rules, improving the efficiency and accuracy of abnormal identification and fully mining the abnormal features in the time-series operation data. Comprehensive time-domain feature indicators are extracted from each time-domain waveform in the state waveform data. The time-domain waveforms are further transformed into frequency-domain waveforms to obtain the corresponding frequency-domain waveforms. Features in the frequency-domain waveforms are mined to obtain the frequency-domain indicator features corresponding to the gearbox and bearings. The abnormal feature indicators extracted from SCADA data and the time-domain and frequency-domain feature indicators extracted from CMS data are concatenated to obtain fused features. Basic information and operating parameters of the unit can also be obtained, better adapting to the actual operating status and real-time data of the unit. The fused features, along with the basic information and operating parameters, are input into a pre-trained fault identification model, which outputs specific fault identification results, improving the efficiency and accuracy of fault identification.

[0191] To implement the above embodiments, this application also proposes an electronic device, including: a processor and a memory communicatively connected to the processor; the memory stores computer execution instructions; the processor executes the computer execution instructions stored in the memory to implement the method provided in the foregoing embodiments.

[0192] To implement the above embodiments, this application also proposes a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the methods provided in the foregoing embodiments.

[0193] To implement the above embodiments, this application also proposes a computer program product, including a computer program that, when executed by a processor, implements the methods provided in the foregoing embodiments.

[0194] The collection, storage, use, processing, transmission, provision, and disclosure of user personal information involved in this application all comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0195] It should be noted that personal information collected from users should be used for legitimate and reasonable purposes and should not be shared or sold outside of these legitimate uses. Furthermore, such collection / sharing should only be conducted after receiving the user's informed consent, including but not limited to notifying the user to read the user agreement / user notice and sign an agreement / authorization that includes authorization of relevant user information before the user uses the function. In addition, any necessary steps must be taken to protect and safeguard access to such personal information data and ensure that others with access to personal information data comply with their privacy policies and procedures.

[0196] This application is intended to provide an implementation scheme for users to selectively prevent the use or access to their personal information data. Specifically, this disclosure is intended to provide hardware and / or software to prevent or block access to such personal information data. Once personal information data is no longer needed, risks can be minimized by restricting data collection and deleting data. Furthermore, where applicable, such personal information is de-identified to protect user privacy.

[0197] In the foregoing descriptions of the embodiments, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0198] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "multiple" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0199] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0200] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0201] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0202] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0203] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0204] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A wind turbine generator train fault detection method, characterized by, The method comprises the following steps: acquiring time sequence operation data and state waveform data of a wind turbine transmission chain; determining an abnormal characteristic index of the wind turbine transmission chain according to whether the time sequence operation data meets a preset target rule; determining a time domain characteristic index and a frequency domain characteristic index of the wind turbine transmission chain according to the state waveform data; determining a fusion feature based on the abnormal characteristic index, the time domain characteristic index and the frequency domain characteristic index, and determining a fault detection result of the wind turbine transmission chain according to the fusion feature; wherein the state waveform data comprises at least one of a main shaft impeller side bearing radial acceleration time domain waveform, a main shaft motor side bearing radial acceleration time domain waveform, a gear box first stage inner ring gear radial acceleration time domain waveform, a gear box low speed shaft radial acceleration time domain waveform, a gear box high speed shaft radial acceleration time domain waveform, a generator driving end radial acceleration time domain waveform and a generator non-driving end radial acceleration time domain waveform; the determination of the time domain characteristic index and the frequency domain characteristic index of the wind turbine transmission chain according to the state waveform data comprises: for any time domain waveform in the state waveform data, acquiring an effective value, a variance, a square root amplitude, a skewness index, a kurtosis index, a kurtosis factor, a waveform index, a peak value index, a margin index, a peak value, a peak-to-peak value and a time domain entropy of the time domain waveform as a time domain characteristic index; converting the time domain waveform into a frequency domain waveform, and determining a frequency domain entropy, an upper envelope line and a corresponding unit component of the frequency domain waveform, the unit component comprising a bearing and a gear box; determining an envelope spectrum entropy according to the upper envelope line; in response to the frequency domain waveform belonging to a bearing, acquiring an inner ring envelope spectrum peak value factor, an outer ring envelope spectrum peak value factor, a rolling element envelope spectrum peak value factor and a retainer envelope spectrum peak value factor based on the upper envelope line, and combining the envelope spectrum entropy and the frequency domain entropy to form a frequency domain characteristic index; in response to the frequency domain waveform belonging to a gear box, acquiring a first stage meshing low speed side frequency energy, a first stage meshing high speed side frequency energy, a second stage meshing low speed side frequency energy, a second stage meshing high speed side frequency energy, a third stage meshing low speed side frequency energy and a third stage meshing high speed side frequency energy of the frequency domain waveform, and combining the envelope spectrum entropy and the frequency domain entropy to form a frequency domain characteristic index; the determination of the fusion feature based on the abnormal characteristic index, the time domain characteristic index and the frequency domain characteristic index, and the determination of the fault detection result of the wind turbine transmission chain according to the fusion feature, comprises: splicing the abnormal characteristic index, the time domain characteristic index and the frequency domain characteristic index to obtain the fusion feature; inputting the fusion feature into a pre-trained fault detection model to obtain a fault detection result output by the fault detection model, the fault detection result comprising a main bearing damage, a generator bearing damage, a gear box tooth surface damage, a gear box parallel stage bearing damage, a poor centering and no fault.

2. The method of claim 1, wherein, The timing operation data includes at least one of the following: generator active power, fan wind speed, generator rotating speed, main shaft rotating speed, unit operation state, generator driving end bearing temperature, generator non-driving end bearing temperature, gear box oil temperature, frequency converter power, gear box driving end bearing temperature, and gear box non-driving end bearing temperature.

3. The method of claim 2, wherein, The method for determining the abnormal characteristic index of the wind turbine generator transmission chain according to whether the timing operation data meets the preset target rule comprises: In response to the unit operation state being grid-connected state, for the same time, based on the generator active power and the generator non-driving end bearing temperature, it is determined whether the generator non-driving end bearing temperature abnormal rule is met to obtain a first abnormal value; In response to the unit operation state being grid-connected state, for the same time, based on the generator active power and the generator driving end bearing temperature, it is determined whether the generator driving end bearing temperature abnormal rule is met to obtain a second abnormal value; In response to the unit operation state being grid-connected state, for the same time, based on the generator active power and the gear box oil temperature, it is determined whether the gear box oil temperature abnormal rule is met to obtain a third abnormal value; According to the frequency converter power and the gear box driving end bearing temperature, it is determined whether the gear box driving end bearing temperature abnormal rule is met to obtain a fourth abnormal value; According to the frequency converter power and the gear box non-driving end bearing temperature, it is determined whether the gear box non-driving end bearing temperature abnormal rule is met to obtain a fifth abnormal value; According to the start-up time length and the generator driving end bearing temperature variation degree, it is determined whether the generator driving end bearing temperature rises too fast abnormal rule is met to obtain a sixth abnormal value; According to the start-up time length and the generator non-driving end bearing temperature variation degree, it is determined whether the generator non-driving end bearing temperature rises too fast abnormal rule is met to obtain a seventh abnormal value; Based on the first abnormal value to the seventh abnormal value, the abnormal characteristic index of the wind turbine generator transmission chain is determined.

4. The method of claim 3, wherein, The method for obtaining the abnormal characteristic index comprises: According to the generator active power, the units of the wind farm are divided into power intervals to obtain a first power interval and a second power interval; For the first power interval, a first continuous time length in which the generator non-driving end bearing temperature of the unit is at least higher than the generator driving end bearing temperature by a fixed value, a second continuous time length in which the generator driving end bearing temperature of the unit is at least higher than the generator non-driving end bearing temperature by a fixed value, and a third continuous time length in which the gear box oil temperature of the unit is higher than a first preset temperature are determined; In response to the first continuous time length being greater than a preset time length, the first abnormal value of the unit is determined to be 1; in response to the second continuous time length being greater than a preset time length, the second abnormal value of the unit is determined to be 1; in response to the third continuous time length being greater than a preset time length, the third abnormal value of the unit is determined to be 1; For the second power interval, in response to the generator non-drive end bearing temperature of the unit being greater than a second preset temperature, a first abnormal value of the unit is determined to be 1; in response to the generator drive end bearing temperature of the unit being greater than the second preset temperature, a second abnormal value of the unit is determined to be 1; in response to the gear box oil temperature of the unit being greater than the second preset temperature, a third abnormal value of the unit is determined to be 1; According to the frequency converter power, it is determined whether the frequency converter power of the unit is greater than a preset power, and in response to the frequency converter power being greater than the preset power, a fourth continuous time length during which the gear box drive end bearing temperature is greater than a third preset temperature and a fifth continuous time length during which the gear box non-drive end bearing temperature is greater than the third preset temperature are respectively determined; In response to the fourth continuous time length being greater than a preset time length, a fourth abnormal value of the unit is determined to be 1, and in response to the fifth continuous time length being greater than the preset time length, a fifth abnormal value of the unit is determined to be 1; According to the starting time length, the unit is divided into a starting initial stage and a normal operation stage; For the unit in the starting initial stage, a first change value of the generator drive end bearing temperature in a fixed time period and a second change value of the generator non-drive end bearing temperature in the fixed time period are determined; in response to the first change values of two consecutive fixed time periods being greater than a first preset change value, a sixth abnormal value of the unit is determined to be 1; in response to the second change values of two consecutive fixed time periods being greater than the first preset change value, a seventh abnormal value of the unit is determined to be 1; For the unit in the normal operation stage, a third change value of the generator drive end bearing temperature in a fixed time period and a fourth change value of the generator non-drive end bearing temperature in the fixed time period are determined; in response to the third change values of two consecutive fixed time periods being greater than a second preset change value, the sixth abnormal value of the unit is determined to be 1; in response to the fourth change values of two consecutive fixed time periods being greater than the second preset change value, the seventh abnormal value of the unit is determined to be 1.

5. The method of claim 1, wherein, The method further comprises: obtaining basic information and working condition parameters of the wind turbine, the basic information including one or more of a power plant, a wind turbine, a measurement point, a rated speed, a vibration waveform global code, a sampling point number, and a time; and the working condition parameters including one or more of a generator speed, a wind speed, an active power, and a main shaft speed; inputting the basic information, the working condition parameters, and the fusion features into a pre-trained fault detection model to obtain a fault detection result of the wind turbine transmission chain.

6. A wind turbine generator transmission chain fault detection apparatus characterized by, The wind turbine transmission chain fault detection method is applied to any one of claims 1-5, comprising: a first obtaining module configured to obtain time sequence operation data and state waveform data of a wind turbine transmission chain; a second obtaining module configured to determine abnormal feature indexes of the wind turbine transmission chain according to whether the time sequence operation data meets a preset target rule; a third obtaining module configured to determine time domain feature indexes and frequency domain feature indexes of the wind turbine transmission chain according to the state waveform data; The fault detection module is configured to determine a fusion feature based on the abnormal feature index, the time domain feature index, and the frequency domain feature index, and determine a fault detection result of the transmission chain of the wind turbine generator set according to the fusion feature.

Images