Method and system for predicting residual life of planet wheel bearing of megawatt wind power speed increaser

Abstract

The invention relates to the technical field of wind power equipment state monitoring and fault prediction, and discloses a method and a system for predicting the residual life of a planet wheel bearing of a megawatt wind power speed increaser. The method comprises the following steps: acquiring vibration acceleration signals of a planet wheel bearing in a plurality of typical load intervals; extracting an envelope spectrum amplitude matched with the inherent fault characteristic frequency in each signal through envelope spectrum analysis; the amplitudes are normalized and then weighted and fused into a comprehensive state index; the index is compared with a threshold value determined based on historical data statistics, and when the threshold value is continuously exceeded, it is judged that an accelerated degradation stage is entered; and performing quadratic polynomial fitting on the index sequence at the accelerated degradation stage by adopting a least square method, and extrapolating and predicting the time when the index reaches a failure threshold as the residual life. According to the method, the problems of inaccurate fault feature extraction, degeneration stage recognition lagging and large life prediction deviation under the variable load working condition are solved, and the prediction precision and the operation and maintenance efficiency are improved.

Description

Technical Field

[0001] This application relates to the field of wind power equipment condition monitoring and fault prediction technology, and in particular to a method and system for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers. Background Technology

[0002] As the core transmission component of a wind turbine, the reliability of the gearbox directly affects the stability and economic benefits of the entire wind farm. The planetary gear system, a critical transmission link within the gearbox, operates under complex conditions of high torque and variable loads for extended periods, making it one of the components with the highest failure rate in wind power equipment.

[0003] Currently, the prediction of the life of planetary gear bearings in engineering practice faces three main technical challenges: First, under variable load conditions, bearing vibration signals are affected by various factors, including instantaneous overload caused by gusts of wind, gear meshing vibration transmission, and environmental noise, resulting in low signal-to-noise ratios of early fault characteristic signals, making it difficult for traditional spectral analysis methods to effectively extract weak fault features. Second, most existing life prediction methods are based on models established under single load conditions or simplified operating conditions, failing to fully consider the differentiated impact of different load levels on the bearing degradation process, and thus failing to accurately reflect the degradation patterns under actual variable load operating conditions. Third, commonly used linear degradation models do not conform to the actual physical mechanisms of bearing failure. Bearing failure processes such as wear and fatigue spalling inherently possess accelerated degradation characteristics, and linear models cannot accurately characterize this nonlinear degradation trend, leading to significant deviations in remaining life prediction results.

[0004] Furthermore, in engineering practice, the setting of state thresholds often relies on engineers' experience and lacks sufficient statistical basis. This makes the identification of accelerated degradation stages highly subjective, easily leading to misjudgments or omissions. Existing prediction methods still fall short of meeting the actual needs of preventive maintenance in wind farms in terms of accuracy and reliability. There is an urgent need to develop a life prediction method that can adapt to variable load conditions, accurately reflect the nonlinear degradation law of bearings, and has sufficient statistical basis. Summary of the Invention

[0005] In order to overcome the above-mentioned defects of the prior art, the embodiments of this application provide a method for predicting the remaining life of planetary gear bearings of megawatt-level wind power speed increasers to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, this application provides a method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers, comprising: Obtain vibration acceleration signals and corresponding timestamp information of planetary gear bearings under multiple preset typical load ranges; Envelope spectrum analysis was performed on the vibration acceleration signal under each typical load range to extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing. The extracted multiple envelope spectrum amplitudes are combined into a comprehensive state index; The comprehensive status index is compared with a preset status index threshold, which is determined based on the critical value that marks the starting point of the accelerated degradation stage in the historical full life data of similar planetary gear bearings. When the comprehensive status index continues to exceed the status index threshold, the planetary gear bearing is determined to have entered the accelerated degradation stage. During the accelerated degradation phase, a time-series-based curve fitting extrapolation method is used to predict the time required for the comprehensive state index to reach a preset failure threshold, which is then used as the remaining service life of the planetary gear bearing.

[0007] To address the aforementioned issues, this application also provides a megawatt-level wind turbine speed increaser planetary gear bearing remaining life prediction system, the system comprising: The signal and timestamp acquisition module is used to acquire the vibration acceleration signals of the planetary gear bearing under multiple preset typical load ranges and the corresponding timestamp information. The envelope spectrum analysis module is used to perform envelope spectrum analysis on the vibration acceleration signal under each typical load range, and extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing. The comprehensive state index combination module is used to combine the extracted multiple envelope spectrum amplitudes into a comprehensive state index. The indicator threshold comparison module is used to compare the comprehensive status indicator with the preset status indicator threshold, which is determined based on the critical value that identifies the starting point of the accelerated degradation stage in the historical full life data of similar bearings of the planetary gear bearing. An accelerated degradation determination module is used to determine that the planetary gear bearing has entered the accelerated degradation stage when the comprehensive status index continuously exceeds the status index threshold. The remaining service life prediction module is used to predict the time required for the comprehensive status index to reach a preset failure threshold during the accelerated degradation stage using a time series-based curve fitting extrapolation method, so as to determine the remaining service life of the planetary gear bearing.

[0008] Compared with the prior art, this application has the following beneficial effects: This application effectively solves the technical challenge of predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers under variable load conditions through multi-load interval fusion analysis and a data-driven threshold setting mechanism. First, by dividing typical load ranges and collecting vibration signals under stable operating conditions, the impact of instantaneous load fluctuations on signal quality is significantly reduced. Then, based on envelope spectrum analysis, the inherent fault characteristic frequency amplitude is extracted, which effectively enhances the signal-to-noise ratio of early fault characteristics. Through normalization processing and weight allocation based on radial force calculation, a comprehensive index that can fully characterize the degradation state of bearings under different load conditions is constructed, overcoming the limitations of single load characterization. This application establishes an objective threshold setting method based on historical full-life data statistics. By using sliding window regression slope analysis to identify the starting point of accelerated degradation, it completely changes the traditional subjective mode of setting thresholds based on experience, and significantly improves the accuracy and reliability of degradation stage identification. Specifically, this method determines the threshold of state indicators with significant statistical significance through statistical analysis of multiple sets of full-life-cycle monitoring data of similar bearings. By combining moving average filtering and regression slope calculation, it effectively identifies the first appearance of a continuous monotonically increasing trend in the comprehensive state indicator sequence, providing an objective basis for accurately judging whether the bearing has entered the accelerated degradation stage. In the life prediction stage, this application abandons the traditional linear prediction model and innovatively adopts a quadratic polynomial to fit the nonlinear trend of the accelerated degradation stage of the bearing, which is more in line with the physical law of accelerated development of bearing wear, fatigue and other failure processes. By solving the coefficients of the fitting function by the least squares method, the optimal balance between accuracy and generalization ability of the model is ensured. This method can significantly reduce the error of remaining life prediction, provide reliable technical support for wind farms to formulate accurate preventive maintenance plans, and effectively avoid unplanned downtime caused by sudden bearing failure. Attached Figure Description

[0009] Figure 1 A flowchart illustrating a method for predicting the remaining life of planetary gear bearings in a megawatt-level wind power speed increaser, as provided in an embodiment of this application. Figure 2 A functional block diagram of a megawatt-level wind turbine speed increaser planetary gear bearing remaining life prediction system provided in an embodiment of this application; The realization of the purpose, functional features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0010] It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit this application.

[0011] This application provides a method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed accelerators. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed accelerators can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms.

[0012] Reference Figure 1 The diagram shown is a flowchart illustrating a method for predicting the remaining life of planetary gear bearings in a megawatt-level wind turbine speed increaser according to an embodiment of this application. In this embodiment, the method for predicting the remaining life of planetary gear bearings in a megawatt-level wind turbine speed increaser includes: S1. Obtain the vibration acceleration signals of the planetary gear bearing under multiple preset typical load ranges and the corresponding timestamp information.

[0013] In this embodiment, the typical load range is a stable operating load range divided according to the rated power of the megawatt-level wind turbine and the degree of influence of the load on bearing degradation. This application sets three ranges, corresponding to 0% to 30%, 30% to 70%, and 70% to 100% of the rated power, respectively. The vibration acceleration signal is a vibration physical quantity that reflects the operating state of the planetary gear bearing, with the unit being m / s², and can carry characteristic information related to bearing failure. The timestamp information is the time data that records the start time of vibration acceleration signal acquisition.

[0014] In some embodiments, acquiring the vibration acceleration signals of the planetary gear bearing under multiple preset typical load ranges and the corresponding timestamp information includes: The instantaneous load of the wind turbine will be compared with several preset typical load ranges; When the instantaneous load falls into a typical load range and remains stable for more than a first preset time, it is determined that the planetary gear bearing has entered the stable operating state corresponding to the typical load range; After determining that the system has entered the stable operating state, the vibration acceleration signal of the planetary gear bearing is collected, and the timestamp information at the start of data collection is recorded simultaneously to form vibration acceleration signal data with timestamps.

[0015] In this embodiment, the hardware is first deployed, including an ICP-type accelerometer, a data acquisition card, and a SCADA system linked to the wind turbine. The ICP-type accelerometer is model PCB352C33, with a frequency range of 0.5 to 10 kHz and a sampling frequency of 25.6 kHz. This sampling frequency setting is based on satisfying the sampling theorem to avoid frequency aliasing and ensure complete acquisition of bearing fault characteristic signals. The data acquisition card supports synchronous data acquisition with the accelerometer. The SCADA system is used to acquire the instantaneous power data of the wind turbine in real time, with an acquisition interval of once per second.

[0016] In this application embodiment, the first step is to acquire and convert instantaneous power data. In this application, the rated power of the megawatt-level wind turbine is set to 2.5MW. After acquiring the instantaneous power from the SCADA system, it is converted into a percentage relative to the rated power. For example, if the instantaneous power acquired at a certain moment is 0.75MW, the corresponding power percentage is 30%, and the instantaneous power acquired at another moment is 1.75MW, the corresponding power percentage is 70%.

[0017] In this embodiment, the second step is to perform a load range comparison, comparing the converted instantaneous power percentage with three preset typical load ranges to determine the range in which the instantaneous power percentage falls. If the instantaneous power percentage is 30%, it falls into the low load range of 0 to 30%; if the instantaneous power percentage is 70%, it falls into the high load range of 70% to 100%; and if the instantaneous power percentage is 50%, it falls into the medium load range of 30% to 70%.

[0018] In this embodiment, the third step is to determine the stable operating state. The criterion for determining the stable operating state is that the fluctuation of the instantaneous power percentage within the corresponding range does not exceed ±5%, and the duration exceeds the first preset time. In this application, the first preset time is set to 10 minutes. The time setting is based on the wind turbine load response curve. The curve shows that the load fluctuation does not exceed ±5% within 10 minutes and can be regarded as stable operation. For example, when the instantaneous power percentage is maintained between 70% and 75% for 10 minutes and the fluctuation does not exceed ±5%, it is determined that the bearing has entered the stable operating state corresponding to the high load range.

[0019] In this embodiment, the fourth step is to acquire signals and record timestamps. After determining that the system has entered a stable operating state, the accelerometer is activated to acquire vibration acceleration signals. The acquisition duration is set to 10 seconds. This duration is set to ensure that it includes 5 bearing rotation cycles. In this application, the bearing speed is 50 r / s, and the duration of 5 rotation cycles is 0.1 seconds. The 10-second acquisition duration can fully cover the fault characteristics. During the acquisition process, the timestamp information at the start of data acquisition is recorded synchronously, and finally, vibration acceleration signal data with timestamps in the format of "load range - vibration acceleration signal - timestamp" is formed.

[0020] In this embodiment of the application, this step avoids signal noise caused by unstable loads such as gusts by dividing typical load ranges and collecting signals under stable conditions, thus solving the problem of large vibration signal interference and inaccurate bearing status information under variable load conditions.

[0021] S2. Perform envelope spectrum analysis on the vibration acceleration signal under each typical load range, and extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing.

[0022] In this embodiment, envelope spectrum analysis is a signal processing method that extracts the envelope of the modulation fault characteristics in the vibration signal and then converts the envelope signal into a frequency domain spectrum, thereby highlighting the analysis method of bearing fault frequency. The inherent fault characteristic frequency is determined by the structural parameters and operating speed of the planetary gear bearing and can reflect the characteristic frequencies of faults in different parts of the bearing, including three types: outer ring fault frequency, inner ring fault frequency, and rolling element fault frequency. The envelope spectrum amplitude is the height of the spectral peak corresponding to the inherent fault characteristic frequency in the envelope spectrum, with the unit being m / s², and the value can directly reflect the severity of the bearing fault.

[0023] In some embodiments, performing envelope spectrum analysis on the vibration acceleration signal under each typical load range and extracting the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing includes: The vibration acceleration signal data with timestamps is subjected to bandpass filtering to obtain the filtered vibration signal. The center frequency of the passband of the bandpass filter is set according to the natural frequency of the planetary gear bearing. Based on the Hilbert transform, the analytic signal of the filtered vibration signal is calculated and the envelope of the analytic signal is extracted to obtain the envelope signal; The envelope signal is converted from the time domain to the frequency domain using a fast Fourier transform to obtain the corresponding envelope spectrum; In the envelope spectrum, spectral peaks corresponding to the outer ring fault characteristic frequency, inner ring fault characteristic frequency, and rolling element fault characteristic frequency, which were calculated in advance based on the structural parameters and operating speed of the planetary gear bearing, were identified. The amplitude of the identified spectral peak is extracted and used as the envelope spectral amplitude that matches the inherent fault characteristic frequency of the planetary gear bearing.

[0024] In this embodiment, the first step is to perform bandpass filtering. The passband center frequency of the bandpass filter is set according to the natural frequency of the planetary gear bearing. In this application, the natural frequency of the bearing is 2kHz, which is obtained from the modal test report provided by the bearing manufacturer. The passband width is set to ±25% of the natural frequency, i.e., 1.5kHz to 2.5kHz. This width setting is based on filtering out power frequency (50Hz) and high-frequency noise. The filtering algorithm uses an FIR filter with a filter order set to 1024 and a stopband attenuation set to 80dB to ensure the filtering effect. For example, the time-stamped vibration acceleration signal [(0.1,0.2,...,2.5)m / s²] in the high load range is filtered to output the filtered vibration signal.

[0025] In this embodiment, the second step is to calculate the analytic signal and extract the envelope. The analytic signal of the filtered vibration signal is calculated based on the Hilbert transform. The calculation process of the Hilbert transform is that the analytic signal equals the filtered vibration signal plus the imaginary unit multiplied by the Hilbert integral of the vibration signal. The Hilbert integral is used to obtain the orthogonal components of the signal through integration operations, for example, for the filtered vibration signal. Its analytical signal The formula for calculation is: ,in, It is the imaginary unit. Indicates the signal Perform a Hilbert transform; after calculating the analytic signal, extract the envelope of the analytic signal. The envelope is obtained by taking the magnitude of the analytic signal, i.e., the envelope signal. For a complex number Its modulus Envelope signal It is a purely real signal that describes the original vibration signal. The amplitude varies with time, for example, the analytical signal is calculated from the filtered signal in the high load region, and then the modulus value is taken to obtain the envelope signal [(0.3,0.5,...,5.2)m / s²].

[0026] In this embodiment, the third step is to convert the envelope signal into an envelope spectrum. The envelope signal is converted from the time domain to the frequency domain using a Fast Fourier Transform (FFT). The number of points in the FFT is set to 1024. This setting is based on ensuring that the frequency resolution meets the requirements. The frequency resolution is equal to the sampling frequency divided by the number of FFT points. In this application, the sampling frequency is 25.6kHz, therefore the frequency resolution is 25.6kHz / 1024 = 25Hz, which can accurately identify the inherent fault characteristic frequency. The envelope spectrum containing the frequency-amplitude correspondence is obtained through transformation. For example, the envelope signal in the high-load range is transformed to obtain the envelope spectrum.

[0027] In this embodiment, the fourth step is to calculate the inherent fault characteristic frequency. The inherent fault characteristic frequency is calculated based on the bearing structural parameters and operating speed. In this application, the bearing model is SKF23030CC / W33, and its structural parameters are derived from the product manual for this bearing model. The specific parameters are: 12 rolling elements, bearing speed 50 r / s, rolling element diameter 50 mm, bearing pitch circle diameter 300 mm, and contact angle 15 degrees. The formula for calculating the outer ring fault frequency is: ,in, For the number of rolling elements, This refers to the bearing rotational speed (revolutions per second). Contact angle, The diameter of the rolling element, Given the bearing pitch circle diameter, substituting the parameters, the calculated outer ring failure frequency is: The formula for calculating the inner ring failure frequency is: Substituting the parameters, the calculated inner ring failure frequency is: The formula for calculating the rolling element failure frequency is: Substituting the parameters, the calculated rolling element failure frequency is: .

[0028] In this embodiment, the fifth step is to identify spectral peaks and extract their amplitudes. In the envelope spectrum, spectral peaks corresponding to the calculated outer ring fault frequency, inner ring fault frequency, and rolling element fault frequency are identified. The threshold for peak identification is set to 30% of the amplitude of the largest peak in the envelope spectrum. For example, if the amplitude of the largest peak in the high-load range envelope spectrum is 5.2 m / s², the identification threshold is 1.56 m / s². Spectral peaks with frequencies close to 251.5 Hz, 348.5 Hz, and 146.1 Hz and amplitudes exceeding 1.56 m / s² are found in the envelope spectrum. For example, a peak with an amplitude of 5.2 m / s² is found near 251 Hz. The spectral peaks were found to be m / s², with an amplitude of 4.8 m / s² around 348 Hz and an amplitude of 4.5 m / s² around 146 Hz. The amplitudes of the identified spectral peaks were extracted, and the amplitude of the largest peak was taken as the envelope spectral amplitude that matches the inherent fault characteristic frequency under the typical load range. For example, the envelope spectral amplitude extracted in the high load range was 5.2 m / s², the envelope spectral amplitude extracted in the low load range (corresponding to an instantaneous power of 0.75 MW) was 2.5 m / s², and the envelope spectral amplitude extracted in the medium load range (corresponding to an instantaneous power of 1.25 MW) was 3.8 m / s².

[0029] In this embodiment, this step accurately extracts the amplitude corresponding to the inherent fault characteristic frequency through envelope spectrum analysis, solving the problem that early fault characteristics are masked by noise and difficult to identify, and providing accurate feature data for the subsequent construction of comprehensive status indicators.

[0030] S3. Combine the extracted envelope spectrum amplitudes into a comprehensive state index.

[0031] In some embodiments, combining the extracted multiple envelope spectrum amplitudes into a comprehensive state index includes: The envelope spectrum amplitudes of different typical load ranges are normalized to obtain the normalized amplitudes corresponding to each envelope spectrum amplitude. A weighting coefficient is assigned to each typical load range, and the magnitude of the weighting coefficient is positively correlated with the importance of the corresponding typical load range in characterizing the degree of bearing degradation. The comprehensive state index is obtained by weighting and summing the weight coefficients of each normalized amplitude and its corresponding typical load range.

[0032] In this embodiment, the comprehensive state index is a quantitative index that integrates the envelope spectrum amplitudes under multiple typical load ranges. It can comprehensively reflect the overall degradation state of the planetary gear bearing, and its value range is 0 to 1 after normalization. Normalization is a method of converting the envelope spectrum amplitudes of different typical load ranges to the same numerical range, with the aim of eliminating the influence of amplitude magnitude differences under different loads on the combined result. The weighting coefficient is a coefficient set according to the importance of each typical load range in representing the degree of bearing degradation. The higher the importance, the larger the weighting coefficient, and the sum of all weighting coefficients is 1.

[0033] In this embodiment, the first step is to normalize the envelope spectrum amplitude. The normalization process uses the minimum-maximum normalization method to linearly map the original envelope spectrum amplitude to the [0,1] interval. The calculation formula for the normalization process is as follows: ,in, For normalized amplitude, The original envelope spectrum amplitude. This represents the minimum envelope spectral amplitude of similar bearings under normal conditions. To determine the maximum envelope spectrum amplitude of similar bearings before failure, based on statistical analysis of the full lifespan data of 10 sets of similar bearings, the following values ​​were set: the minimum envelope spectrum amplitude under normal conditions for similar bearings is 1.0 m / s², and the maximum envelope spectrum amplitude before failure is 20.0 m / s². These values ​​are based on the statistical analysis of the full lifespan data of 10 sets of similar bearings; under normal conditions, the amplitudes are all greater than 1.0 m / s², and before failure, the amplitudes are all close to 20.0 m / s². For example, the envelope spectrum amplitude of 2.5 m / s² in the low load range was analyzed. After normalization, the normalized amplitude is calculated as (2.5-1.0) / (20.0-1.0)=1.5 / 19≈0.0789; for the envelope spectrum amplitude of 3.8 m / s² in the medium load range, the normalized amplitude is calculated as (3.8-1.0) / 19=2.8 / 19≈0.1474; for the envelope spectrum amplitude of 5.2 m / s² in the high load range, the normalized amplitude is calculated as (5.2-1.0) / 19=4.2 / 19≈0.2211.

[0034] In this embodiment, the second step is to set weighting coefficients. The weighting coefficients are set based on the importance of each load range in representing the degree of bearing degradation. The fundamental basis is the magnitude of the radial force on the bearing under different load ranges. The larger the radial force, the more important the vibration signal is in representing the degradation process. The formula for calculating the radial force is 9550 multiplied by the power and then divided by the product of the rotational speed and the bearing pitch circle diameter, where the power unit is kilowatts, the rotational speed unit is revolutions per minute, and the diameter unit is meters. For the high load range, the power is 2500 kilowatts (2.5MW), the rotational speed is 3000 revolutions per minute (50r / s), and the bearing pitch circle diameter is 0.3 meters. The calculated radial force is (9550 × 10⁻⁶ m / s). 50×2500) / (3000×0.3)≈26527.8 N; For the medium load range, the power is 1250 kW (1.25 MW), and the calculated radial force is 13263.9 N; For the low load range, the power is 750 kW (0.75 MW), and the calculated radial force is 7958.3 N; The radial force in the high load range is about 3.3 times that in the low load range, and the radial force in the medium load range is about 1.67 times that in the low load range. Therefore, the weighting coefficient for the high load range is set to 0.5, the weighting coefficient for the medium load range is 0.35, and the weighting coefficient for the low load range is 0.15. The sum of the weighting coefficients is 0.5+0.35+0.15=1.

[0035] In this embodiment of the application, the third step is to calculate the comprehensive state index. The comprehensive state index is obtained by weighted summation of the normalized amplitude of each load interval and its corresponding weight coefficient. For example, the comprehensive state index calculated by substituting the values ​​from the above steps is 0.0789×0.15+0.1474×0.35+0.2211×0.5≈0.174, where the contribution of low load is 0.0789×0.15≈0.0118, the contribution of medium load is 0.1474×0.35≈0.0516, and the contribution of high load is 0.2211×0.50≈0.1106.

[0036] In this embodiment of the application, this step constructs a comprehensive state index by normalization and weighted summation, which eliminates the influence of amplitude differences under different loads, comprehensively reflects the bearing degradation state, solves the problem that the index under a single load cannot fully characterize the bearing state, and provides a reliable basis for subsequent degradation stage determination.

[0037] S4. Compare the comprehensive status index with the preset status index threshold, which is determined based on the critical value that marks the starting point of the accelerated degradation stage in the historical full life data of similar planetary gear bearings.

[0038] In this embodiment, the state index threshold is a critical comprehensive state index value that distinguishes between the normal degradation stage and the accelerated degradation stage of a planetary gear bearing. When the comprehensive state index exceeds this value, it indicates that the bearing may have entered the accelerated degradation stage. The historical full life data of similar bearings is the complete monitoring data of bearings of the same model and operating conditions as the planetary gear bearing to be monitored, from normal service state to complete failure state, including comprehensive state index values ​​collected at different operating time points. The starting point of the accelerated degradation stage is the starting time point when the bearing degradation rate changes from slow to fast, and the corresponding comprehensive state index value is the critical value that identifies this starting point.

[0039] In some embodiments, the state index threshold is determined based on a critical value identifying the starting point of the accelerated degradation stage in the historical full-life data of similar planetary gear bearings, including: Acquire historical monitoring data of multiple sets of similar bearings throughout their entire life cycle from normal service to complete failure. The historical monitoring data includes comprehensive status index values ​​collected at different operating time points. For each set of historical monitoring data, identify the starting data point in the sequence of comprehensive status index values ​​of the historical monitoring data where a continuous monotonous upward trend first appears, and mark the starting data point as the critical point of the corresponding bearing accelerated degradation stage; Extract the comprehensive state index values ​​at each critical point identified in all historical monitoring data to form a set of critical point index values, and use the arithmetic mean of the set of critical point index values ​​as the state index threshold.

[0040] In some embodiments, identifying the starting data point in the sequence of comprehensive state index values ​​of the historical monitoring data where a sustained monotonically increasing trend first appears for each set of historical monitoring data, and marking the starting data point as the critical point of the corresponding bearing accelerated degradation stage, includes: The smoothed index sequence is obtained by performing a smoothing filter on the comprehensive status index value sequence in each set of historical monitoring data. The smoothed index sequence is traversed using a sliding time window of fixed width. For each window position, the regression slope of the comprehensive state index value relative to time within the window is calculated. When the first sliding window with a regression slope greater than zero is encountered, it is marked as an undetermined critical window; After the undetermined critical window, the regression slope of the subsequent preset number of sliding windows is monitored. If more than half of the subsequent preset number of sliding windows have a regression slope greater than zero, then the continuous monotonically upward trend is determined to be valid, and the starting point of the undetermined critical window is formally identified as the starting data point. The timestamp and comprehensive status index value corresponding to the starting data point are recorded as the characterization data of the critical point.

[0041] In this embodiment, the first step is to acquire historical full-life data of similar bearings. The operating conditions are set as wind speed of 7 to 12 m / s and ambient temperature of -10 to 40°C, which are consistent with the operating conditions of the bearing to be monitored. A total of 10 sets of historical full-life data are acquired, with a monitoring period of 5 minutes for each set of data. The data includes all comprehensive status index values ​​from the bearing installation and use to complete failure. For example, a set of historical data contains more than 5,000 comprehensive status index values, with a time span from 0 hours of bearing operation to 1,500 hours of operation (the moment of complete failure).

[0042] In this embodiment, the second step is to identify the critical point of each set of historical data. This step aims to accurately find the turning point from the stable degradation stage to the accelerated degradation stage of the bearing from the historical data. First, the comprehensive state index value sequence in each set of historical monitoring data is smoothed and filtered using a moving average filtering method. The filtering window size is set to 5 monitoring periods (25 minutes) to eliminate the influence of instantaneous noise on trend identification. For example, a moving average filtering is applied to the comprehensive state index value sequence [0.05, 0.06, 0.05, 0.07, 0.06, ..., 0.8] of a certain set of historical data to obtain the smoothed index sequence [0.058, 0.062, 0.064, ..., 0.79]. Second, a fixed-width sliding time window is used to traverse the smoothed index sequence. The sliding window width is set to 5 monitoring periods (25 minutes), consistent with the smoothing filtering window, to ensure consistency in trend identification. For each window position, the regression slope of the comprehensive state index value relative to time within the window is calculated. The regression slope is calculated as follows: The number of data points within the window (n=5). For a point in time (in hours). This is a comprehensive state index value. The regression slope reflects the trend of change. For example, within a certain window, the time points (hours) are [500, 500.083, 500.166, 500.25, 500.333], and the corresponding comprehensive state index values ​​are [0.14, 0.15, 0.16, 0.15, 0.17]. With 5 data points, a regression slope greater than 0 indicates an upward trend in the index within the window. A two-level verification mechanism ensures the reliability of critical point identification. When the first sliding window with a regression slope greater than 0 is encountered, the index... This is denoted as the undetermined critical window. After the undetermined critical window, the regression slope of the subsequent three sliding windows is monitored. If the regression slope of at least two windows is greater than 0, the continuous monotonous upward trend is determined to be valid. The starting point of the undetermined critical window is identified as the critical point of the accelerated degradation stage. For example, if a set of historical data first shows a regression slope greater than 0 at the window's corresponding time point of 550 hours, and two of the subsequent three windows show a regression slope greater than 0, then the comprehensive state index value of 0.16 corresponding to 550 hours is determined as the critical point index value of this set of data.

[0043] In this embodiment, after constructing a set of 10 historical data critical point index values, the arithmetic mean of the set is calculated to determine the state index threshold. The calculation process for the arithmetic mean is as follows: First, all critical point index values ​​in the set are added together. The 10 values ​​are 0.15, 0.16, 0.17, 0.15, 0.16, 0.17, 0.16, 0.15, 0.17, and 0.16, respectively, and the sum is 1.6. Then, the sum is divided by the number of data sets, 10, to obtain the arithmetic mean, which is 1.6 ÷ 10 = 0.16. This arithmetic mean is the preset state index threshold. The arithmetic mean is chosen as the threshold because it can balance the differences between multiple sets of historical data, avoid threshold deviation caused by single or extreme data, and ensure the statistical significance and universality of the threshold.

[0044] In this embodiment, after setting the state index threshold, the comprehensive state index calculated in step 3 is compared with the threshold. In this application, the comprehensive state index output in step S3 is 0.174, and the state index threshold is 0.16. By comparing the numerical values, 0.174 > 0.16, thus obtaining a preliminary result that the current comprehensive state index exceeds the preset state index threshold. The comparison process uses direct numerical comparison, based on the fact that the comprehensive state index and the threshold have the same dimensions (both are normalized dimensionless values). Direct comparison can intuitively reflect whether the index has reached a critical state without additional conversion or correction.

[0045] In this embodiment of the application, this step determines the threshold based on historical full-life data of similar bearings and compares it, avoiding the judgment deviation caused by subjectively setting the threshold, solving the problem of unreasonable threshold setting and inability to accurately identify the degradation stage, and providing an objective basis for the subsequent determination of the accelerated degradation stage.

[0046] S5. When the comprehensive status index continues to exceed the status index threshold, it is determined that the planetary gear bearing has entered the accelerated degradation stage.

[0047] In some embodiments, determining that the planetary gear bearing has entered an accelerated degradation stage when the comprehensive condition index continuously exceeds the condition index threshold includes: Within multiple consecutive monitoring cycles with equal time intervals, a series of comprehensive status index values ​​arranged in chronological order are collected and calculated to form the current comprehensive status index sequence. Identify whether the starting data point of a sustained monotonically upward trend appears for the first time in the current comprehensive state index sequence; If the starting data point is identified, and starting from the starting data point, a predetermined number of consecutive comprehensive status index values ​​are all greater than the status index threshold, then the planetary gear bearing is determined to have entered the accelerated degradation stage.

[0048] In this embodiment of the application, "continuous exceedance" means that the comprehensive status index is greater than the preset status index threshold in multiple consecutive monitoring cycles with equal time intervals, rather than a single or discontinuous exceedance; the accelerated degradation stage is the operating stage in which the degradation rate of the planetary gear bearing changes from slow to fast. During this stage, the bearing failure will develop at a rate significantly higher than that of the normal degradation stage, which is the key stage for predicting the remaining life.

[0049] In this embodiment, the first step is to determine the criteria for "continuous exceedance," including the number of continuous monitoring cycles and the duration of each monitoring cycle. The duration of the monitoring cycle is set to 5 minutes, which matches the stability determination cycle of signal acquisition in step 1, ensuring consistency in the time dimension between data acquisition and state determination. The number of continuous monitoring cycles is set to 5, i.e., the total duration is 5×5=25 minutes. This number is based on the statistical data of actual wind farm operation. The instantaneous index exceeding the threshold caused by interference factors such as gusts and instantaneous fluctuations in the power grid usually does not exceed 10 minutes in duration. 25 minutes of continuous monitoring can effectively eliminate such interference and avoid misjudgment.

[0050] In this embodiment, the second step is to acquire comprehensive status index data for a continuous monitoring period. Based on the signal acquisition frequency in step S1 and the comprehensive status index calculation method in step S3, a comprehensive status index value is generated every 5 minutes. After the comprehensive status index is first detected to exceed the threshold (0.16) in step S4, data for the next 4 monitoring periods is collected to form a continuous index sequence of 5 periods. For example, the index value that first exceeds the threshold is 0.174 (corresponding to 1000.083 hours), and the index values ​​for the next 4 periods are 0.182 (1000.166 hours), 0.190 (1000.250 hours), 0.185 (1000.333 hours), and 0.195 (1000.416 hours), respectively, forming a continuous index sequence [0.174, 0.182, 0.190, 0.185, 0.195].

[0051] In this embodiment, the third step is to verify whether the indicator sequence continuously exceeds the threshold. The comprehensive state indicator values ​​for five consecutive periods are compared with the state indicator threshold of 0.16, confirming whether each indicator value is greater than the threshold. In the above sequence, 0.174 > 0.16, 0.182 > 0.16, 0.190 > 0.16, 0.185 > 0.16, and 0.195 > 0.16; all indicator values ​​meet the condition of exceeding the threshold.

[0052] In this embodiment, the fourth step is to determine whether the planetary gear bearing has entered the accelerated degradation stage. Since the comprehensive status index exceeds a preset threshold for five consecutive monitoring cycles, and instantaneous interference factors have been excluded, the planetary gear bearing is determined to have entered the accelerated degradation stage. No additional parameters are required during the determination process; verification is based solely on the continuous exceeding-threshold characteristic of the comprehensive status index. When the index continuously exceeds the threshold, the bearing degradation rate will significantly increase in the subsequent time, which conforms to the core characteristics of the accelerated degradation stage.

[0053] In this embodiment of the application, this step effectively eliminates misjudgments caused by instantaneous interference through the judgment logic of continuous cycle verification, accurately identifies the accelerated degradation stage of the bearing, solves the problem of misjudgment of the degradation stage due to failure to distinguish between instantaneous fluctuations and the actual degradation trend, and provides the correct time starting point for subsequent accurate prediction of the remaining life.

[0054] S6. During the accelerated degradation stage, a time series-based curve fitting extrapolation method is used to predict the time required for the comprehensive state index to reach the preset failure threshold, which is used as the remaining service life of the planetary gear bearing.

[0055] In some embodiments, during the accelerated degradation phase, predicting the time required for the comprehensive condition index to reach a preset failure threshold using a time-series-based curve fitting extrapolation method, as the remaining service life of the planetary gear bearing, includes: Multiple comprehensive state index values ​​and their corresponding timestamps, arranged in chronological order since the determination of entering the accelerated degradation stage, are obtained to form a comprehensive state index time series. The least squares fitting algorithm is used to perform curve fitting on the time series of the comprehensive state index to obtain a degradation trend line characterizing the change law of the comprehensive state index over time. Extend the degradation trend line forward, calculate the intersection point of the degradation trend line and the preset failure threshold, and determine the time coordinate corresponding to the intersection point; Calculate the time difference between the current time point and the intersection time coordinate, and use the time difference as the remaining service life of the planetary gear bearing.

[0056] In some embodiments, the step of using a least-squares fitting algorithm to perform curve fitting on the time series of the comprehensive state index to obtain a degradation trend line characterizing the change law of the comprehensive state index over time includes: The timestamps in the comprehensive state index time series are converted into relative time values ​​relative to the starting point of the accelerated degradation stage, and a quadratic polynomial function is used as the fitting function. Based on the least squares principle, the function coefficients that minimize the sum of squared residuals between the quadratic polynomial function value and the actual comprehensive state index value are solved to generate the mathematical expression of the degradation trend line.

[0057] In this embodiment, the curve fitting extrapolation method is a method that uses historical data to construct a mathematical model, fit the trend of data changes, and then extends the trend to a certain critical value in the future, thereby predicting the time required for the target parameter to reach the critical value. The time series of the comprehensive state index is an ordered data set composed of comprehensive state indexes collected and calculated at fixed time intervals during the accelerated degradation stage, together with the corresponding collection time. The remaining life refers to the remaining operating time from the current monitoring time until the comprehensive state index of the planetary gear bearing reaches the preset failure threshold and cannot operate normally. The preset failure threshold is a critical value of the comprehensive state index that characterizes the complete failure of the planetary gear bearing and its inability to continue service. This value is determined based on measured data when similar bearings fail.

[0058] In this embodiment, the first step is to construct a time series of comprehensive state indicators for the accelerated degradation stage. Using the starting point of the accelerated degradation stage determined in step S5 as the time reference, the cumulative running time corresponding to this starting point is set to 1000 hours (i.e., accelerated degradation start time t0 = 1000 hours). The monitoring period remains 5 minutes (converted to hours 5 ÷ 60 ≈ 0.083 hours). Comprehensive state indicator data within the accelerated degradation stage are collected and the corresponding times are recorded. For example, the cumulative running times corresponding to the five sets of collected data are 1000.083 hours, 1000.166 hours, 1000.250 hours, 1000.333 hours, and 1000.416 hours, respectively, and the corresponding comprehensive state indicator values ​​are 0.174, 0.182, 0.190, 0.185, and 0.195, respectively. This constructs a time series dataset, where each data point contains two parameters: "cumulative running time - comprehensive state indicator," providing basic data for subsequent fitting.

[0059] In this embodiment, the second step is to determine the type of fitting function. This step aims to select the most suitable mathematical model for predicting the trend of accelerated degradation in bearings. By analyzing time series data from 10 sets of similar bearings in the accelerated degradation stage, the goodness of fit of different fitting functions (measured by the coefficient of determination R²) was compared. The average R² of linear fitting was approximately 0.85, the average R² of quadratic polynomial fitting was approximately 0.98, and the average R² of cubic polynomial fitting was approximately 0.99, but there was a risk of overfitting (significant prediction deviation for new data). Considering the balance between fitting accuracy and generalization ability, a quadratic polynomial was chosen as the fitting function, and its expression is a comprehensive state index. for: ,in, It is relative time, that is , It is the cumulative running time. This refers to the starting point (critical point) of the accelerated degradation stage identified in the aforementioned steps, where a, b, and c are the coefficients of the quadratic polynomial. The core basis for choosing the quadratic polynomial is that the failure processes such as wear and fatigue in the accelerated degradation stage of bearings exhibit nonlinear characteristics. The quadratic function can more accurately characterize the law that "the degradation rate gradually increases with time," avoiding the defect that linear fitting cannot reflect the rate change.

[0060] In this embodiment, the third step is to solve for the coefficients a, b, and c of the fitted function using the least squares method. First, the cumulative running time is converted to relative time, according to... The relative times for the above five sets of data are 0.083 hours, 0.166 hours, 0.250 hours, 0.333 hours, and 0.416 hours, respectively, and the corresponding comprehensive state index values ​​are still 0.174, 0.182, 0.190, 0.185, and 0.195. The principle of the least squares method is to minimize the approximation between the actual index value and the fitted function. The optimal coefficients are determined by calculating the sum of squared residuals of the calculated values. The core equations are derived based on the mathematical condition of "minimizing the sum of squared residuals," where the sum of squared residuals is... The specific system of equations is as follows: ; Substitute the 5 sets of data into the system of equations for calculation: First, calculate the sum of each term, with the number of data points p=5; the sum of all relative times is 0.083+0.166+0.250+0.333+0.416≈1.248 hours; the sum of the squares of all relative times is 0.083. 2 +0.166 2 +0.250 2 +0.333 2 +0.416 2 ≈0.3809 hours 2 The sum of all relative time cubes is 0.083. 3 +0.166 3 +0.250 3 +0.333 3 +0.416 3 ≈0.1296 hours 3 The sum of all relative times raised to the fourth power is 0.083. 4 +0.166 4 +0.250 4 +0.333 4 +0.416 ≈ 0.0470 hours 4 The sum of all comprehensive state index values ​​is 0.174 + 0.182 + 0.190 + 0.185 + 0.195 = 0.926; the sum of the products of all relative times and their corresponding comprehensive state index values ​​is 0.083 × 0.174 + 0.166 × 0.182 + 0.250 × 0.190 + 0.333 × 0.185 + 0.416 × 0.19 ≈ 0.2368 hours; the sum of the products of the squares of all relative times and their corresponding comprehensive state index values ​​is 0.0069 × 0.174 + 0.0276 × 0.182 + 0.0625 × 0.190 + 0.1109 × 0.185 + 0.1730 × 0.195 ≈ 0.0723 hours. 2 .

[0061] Substituting the summation results into the system of equations, we get: 0.926 = 5c + 1.248b + 0.3809a; 0.2368 = 1.248c + 0.3809b + 0.1296a; 0.0723 = 0.3809c + 0.1296b + 0.0470a.

[0062] By solving the system of linear equations (using matrix inversion or elimination), the coefficients a≈0.02, b≈0.12, and c≈0.16 are calculated. Therefore, the fitted function is a comprehensive state index. .

[0063] In this embodiment, the fourth step is to set a preset failure threshold. The preset failure threshold is determined based on the measured data of the comprehensive state index when the bearings of the same type completely fail. By statistically analyzing the final state index values ​​of 10 groups of failed bearings, the normalized average value is 0.8. Therefore, the preset failure threshold is set to 0.8 (the original envelope spectrum amplitude corresponding to this value can be obtained by the inverse normalization operation, that is, the original amplitude = failure threshold × (maximum amplitude - minimum amplitude) + minimum amplitude = 0.8 × (20.0 - 1.0) + 1.0 = 16.2 m / s², which corresponds to the actual damage degree when the bearing fails (such as the outer ring peeling area ≥ 5 mm²).

[0064] In this embodiment, the fifth step is to extrapolate and calculate the failure time. A preset failure threshold is substituted into the fitting function to solve for the relative time t', i.e., solving the equation 0.02t'² + 0.12t' + 0.16 = 0.8. First, the equation is rearranged into the standard quadratic form 0.02t'² + 0.12t' - 0.64 = 0. The discriminant Δ = 0.12² - 4 × 0.02 × (-0.64) = 0.0656 is calculated using the quadratic formula. Since the relative time is positive, the positive root t' = [-0.12 + 0.256] / (2 × 0.02) ≈ 3.18 hours is taken. This relative time is the time required from the start of accelerated degradation to failure.

[0065] In this embodiment, the sixth step is to calculate the remaining lifetime. The cumulative operating time corresponding to the current monitoring time is 1000.416 hours (i.e., the last cycle time of 5 consecutive monitoring cycles), the accelerated degradation start time t0 = 1000 hours, the failure time is the accelerated degradation start time plus the relative failure time, that is, the failure time is 1000 + 3.18 = 1003.18 hours, and the remaining lifetime is the failure time minus the cumulative operating time at the current monitoring time, that is, the remaining lifetime is 1003.18 - 1000.416 ≈ 2.76 hours. Finally, the remaining lifetime prediction result of the planetary gear bearing is obtained.

[0066] In this embodiment of the application, this step accurately characterizes the nonlinear trend of the accelerated degradation stage by selecting a quadratic polynomial fitting function and solving it using the least squares method. Compared with the traditional linear prediction method, it significantly reduces the prediction error of the remaining lifetime, solves the problem of inaccurate remaining lifetime prediction and inability to guide operation and maintenance decisions, and provides accurate time basis for wind farms to formulate preventive maintenance plans.

[0067] like Figure 2 The diagram shown is a functional block diagram of a megawatt-level wind turbine speed increaser planetary gear bearing remaining life prediction system provided in an embodiment of this application.

[0068] The megawatt-level wind turbine speed increaser planetary gear bearing remaining life prediction system 100 described in this application can be installed in an electronic device. Depending on the functions implemented, the megawatt-level wind turbine speed increaser planetary gear bearing remaining life prediction system 100 may include a signal and timestamp acquisition module 101, an envelope spectrum analysis module 102, a comprehensive state index combination module 103, an index threshold comparison module 104, an accelerated degradation determination module 105, and a remaining life prediction module 106. The module described in this application can also be referred to as a unit, which refers to a series of computer program segments that can be executed by the processor of an electronic device and can perform a fixed function, and are stored in the memory of the electronic device.

[0069] In this embodiment, the functions of each module / unit are as follows: The signal and timestamp acquisition module 101 is used to acquire the vibration acceleration signals of the planetary gear bearing under multiple preset typical load ranges and the corresponding timestamp information. The envelope spectrum analysis module 102 is used to perform envelope spectrum analysis on the vibration acceleration signal under each typical load range, and extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing. The comprehensive state index combination module 103 is used to combine the extracted multiple envelope spectrum amplitudes into a comprehensive state index; The index threshold comparison module 104 is used to compare the comprehensive status index with the preset status index threshold, which is determined based on the critical value that identifies the starting point of the accelerated degradation stage in the historical full life data of the same type of planetary gear bearing. The accelerated degradation determination module 105 is used to determine that the planetary gear bearing has entered the accelerated degradation stage when the comprehensive status index continuously exceeds the status index threshold. The remaining service life prediction module 106 is used to predict the time required for the comprehensive status index to reach the preset failure threshold during the accelerated degradation stage by using a time series-based curve fitting extrapolation method, so as to determine the remaining service life of the planetary gear bearing.

[0070] In the several embodiments provided in this application, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0071] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0072] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0073] It will be apparent to those skilled in the art that this application is not limited to the details of the exemplary embodiments described above, and that this application can be implemented in other specific forms without departing from the spirit or essential characteristics of this application.

[0074] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of this application without departing from the spirit and scope of the technical solutions of this application.

Claims

1. A method for predicting the remaining life of planetary gear bearings in a megawatt-level wind turbine speed increaser, characterized in that, The method includes: Obtain vibration acceleration signals and corresponding timestamp information of planetary gear bearings under multiple preset typical load ranges; Envelope spectrum analysis was performed on the vibration acceleration signal under each typical load range to extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing. The extracted multiple envelope spectrum amplitudes are combined into a comprehensive state index; The comprehensive status index is compared with a preset status index threshold, which is determined based on the critical value that marks the starting point of the accelerated degradation stage in the historical full life data of similar planetary gear bearings. When the comprehensive status index continues to exceed the status index threshold, the planetary gear bearing is determined to have entered the accelerated degradation stage. During the accelerated degradation phase, a time-series-based curve fitting extrapolation method is used to predict the time required for the comprehensive state index to reach a preset failure threshold, which is then used as the remaining service life of the planetary gear bearing.

2. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 1, characterized in that, The acquisition of vibration acceleration signals and corresponding timestamp information of planetary gear bearings under multiple preset typical load ranges includes: The instantaneous load of the wind turbine will be compared with several preset typical load ranges; When the instantaneous load falls into a typical load range and remains stable for more than a first preset time, it is determined that the planetary gear bearing has entered the stable operating state corresponding to the typical load range; After determining that the system has entered the stable operating state, the vibration acceleration signal of the planetary gear bearing is collected, and the timestamp information at the start of data collection is recorded simultaneously to form vibration acceleration signal data with timestamps.

3. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 1, characterized in that, The process of performing envelope spectrum analysis on the vibration acceleration signals under each typical load range, and extracting the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing, includes: The vibration acceleration signal data with timestamps is subjected to bandpass filtering to obtain the filtered vibration signal. The center frequency of the passband of the bandpass filter is set according to the natural frequency of the planetary gear bearing. Based on the Hilbert transform, the analytic signal of the filtered vibration signal is calculated and the envelope of the analytic signal is extracted to obtain the envelope signal; The envelope signal is converted from the time domain to the frequency domain using a fast Fourier transform to obtain the corresponding envelope spectrum; In the envelope spectrum, spectral peaks corresponding to the outer ring fault characteristic frequency, inner ring fault characteristic frequency, and rolling element fault characteristic frequency, which were calculated in advance based on the structural parameters and operating speed of the planetary gear bearing, were identified. The amplitude of the identified spectral peak is extracted and used as the envelope spectral amplitude that matches the inherent fault characteristic frequency of the planetary gear bearing.

4. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 1, characterized in that, The process of combining the extracted multiple envelope spectrum amplitudes into a comprehensive state index includes: The envelope spectrum amplitudes of different typical load ranges are normalized to obtain the normalized amplitudes corresponding to each envelope spectrum amplitude. A weighting coefficient is assigned to each typical load range, and the magnitude of the weighting coefficient is positively correlated with the importance of the corresponding typical load range in characterizing the degree of bearing degradation. The comprehensive state index is obtained by weighting and summing the weight coefficients of each normalized amplitude and its corresponding typical load range.

5. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 1, characterized in that, The state index threshold is determined based on the critical value that identifies the starting point of the accelerated degradation stage in the historical full-life data of similar planetary gear bearings, including: Acquire historical monitoring data of multiple sets of similar bearings throughout their entire life cycle from normal service to complete failure. The historical monitoring data includes comprehensive status index values ​​collected at different operating time points. For each set of historical monitoring data, identify the starting data point in the sequence of comprehensive status index values ​​of the historical monitoring data where a continuous monotonous upward trend first appears, and mark the starting data point as the critical point of the corresponding bearing accelerated degradation stage; Extract the comprehensive state index values ​​at each critical point identified in all historical monitoring data to form a set of critical point index values, and use the arithmetic mean of the set of critical point index values ​​as the state index threshold.

6. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 5, characterized in that, For each set of historical monitoring data, the step of identifying the starting data point in the sequence of comprehensive state index values ​​of the historical monitoring data where a sustained monotonically increasing trend first appears, and marking the starting data point as the critical point of the corresponding bearing accelerated degradation stage, includes: The smoothed index sequence is obtained by performing a smoothing filter on the comprehensive status index value sequence in each set of historical monitoring data. The smoothed index sequence is traversed using a sliding time window of fixed width. For each window position, the regression slope of the comprehensive state index value relative to time within the window is calculated. When the first sliding window with a regression slope greater than zero is encountered, it is marked as an undetermined critical window; After the undetermined critical window, the regression slope of the subsequent preset number of sliding windows is monitored. If more than half of the subsequent preset number of sliding windows have a regression slope greater than zero, then the continuous monotonically upward trend is determined to be valid, and the starting point of the undetermined critical window is formally identified as the starting data point. The timestamp and comprehensive status index value corresponding to the starting data point are recorded as the characterization data of the critical point.

7. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 6, characterized in that, The step of determining that the planetary gear bearing has entered the accelerated degradation stage when the comprehensive condition index continuously exceeds the condition index threshold includes: Within multiple consecutive monitoring cycles with equal time intervals, a series of comprehensive status index values ​​arranged in chronological order are collected and calculated to form the current comprehensive status index sequence. Identify whether the starting data point of a sustained monotonically upward trend appears for the first time in the current comprehensive state index sequence; If the starting data point is identified, and starting from the starting data point, a predetermined number of consecutive comprehensive status index values ​​are all greater than the status index threshold, then the planetary gear bearing is determined to have entered the accelerated degradation stage.

8. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind turbine speed increasers as described in claim 1, characterized in that, During the accelerated degradation phase, a time-series-based curve fitting extrapolation method is used to predict the time required for the comprehensive state index to reach a preset failure threshold, which is then used as the remaining service life of the planetary gear bearing. This includes: Multiple comprehensive state index values ​​and their corresponding timestamps, arranged in chronological order since the determination of entering the accelerated degradation stage, are obtained to form a comprehensive state index time series. The least squares fitting algorithm is used to perform curve fitting on the time series of the comprehensive state index to obtain a degradation trend line characterizing the change law of the comprehensive state index over time. Extend the degradation trend line forward, calculate the intersection point of the degradation trend line and the preset failure threshold, and determine the time coordinate corresponding to the intersection point; Calculate the time difference between the current time point and the intersection time coordinate, and use the time difference as the remaining service life of the planetary gear bearing.

9. The method for predicting the remaining life of planetary gear bearings in megawatt-level wind power speed increasers as described in claim 8, characterized in that, The step of using a least-squares fitting algorithm to perform curve fitting on the time series of the comprehensive state index to obtain a degradation trend line characterizing the change law of the comprehensive state index over time includes: The timestamps in the comprehensive state index time series are converted into relative time values ​​relative to the starting point of the accelerated degradation stage, and a quadratic polynomial function is used as the fitting function. Based on the least squares principle, the function coefficients that minimize the sum of squared residuals between the quadratic polynomial function value and the actual comprehensive state index value are solved to generate the mathematical expression of the degradation trend line.

10. A system for predicting the remaining life of planetary gear bearings in a megawatt-level wind turbine speed increaser, used to implement the method for predicting the remaining life of planetary gear bearings in a megawatt-level wind turbine speed increaser as described in any one of claims 1-9, characterized in that, The system includes: The signal and timestamp acquisition module is used to acquire the vibration acceleration signals of the planetary gear bearing under multiple preset typical load ranges and the corresponding timestamp information. The envelope spectrum analysis module is used to perform envelope spectrum analysis on the vibration acceleration signal under each typical load range, and extract the envelope spectrum amplitude of each signal that matches the inherent fault characteristic frequency of the planetary gear bearing. The comprehensive state index combination module is used to combine the extracted multiple envelope spectrum amplitudes into a comprehensive state index. The indicator threshold comparison module is used to compare the comprehensive status indicator with the preset status indicator threshold, which is determined based on the critical value that identifies the starting point of the accelerated degradation stage in the historical full life data of similar bearings of the planetary gear bearing. An accelerated degradation determination module is used to determine that the planetary gear bearing has entered the accelerated degradation stage when the comprehensive status index continuously exceeds the status index threshold. The remaining service life prediction module is used to predict the time required for the comprehensive status index to reach a preset failure threshold during the accelerated degradation stage using a time series-based curve fitting extrapolation method, so as to determine the remaining service life of the planetary gear bearing.

Images