CN-122020331-A - Wind driven generator bearing state analysis method, system and medium

CN122020331ACN 122020331 ACN122020331 ACN 122020331ACN-122020331-A

Abstract

The application provides a wind driven generator bearing state analysis method, a system and a medium, wherein the method comprises the steps of collecting running parameters of a wind driven generator bearing in real time based on a sensor group to obtain multidimensional running parameter information, carrying out preprocessing operation on the multidimensional running parameter information, respectively extracting time domain features and frequency domain features based on the preprocessed multidimensional running parameter information to obtain feature sets, carrying out dimension reduction optimization processing on the feature sets by adopting a feature selection algorithm to obtain an optimal feature subset, outputting real-time state information of the wind driven generator bearing based on a bearing state recognition model, carrying out state evaluation and abnormal early warning on the real-time state information of the wind driven generator bearing based on a dynamic early warning mechanism to obtain a state analysis result, and carrying out automatic evaluation and recognition on the wind driven generator bearing state to realize direct recognition of the result, so that the technical requirements of wind field maintenance personnel are reduced, high-density equipment state evaluation is realized, and the reliability of equipment operation and evaluation is improved.

Inventors

WANG YONG
Ye Linglin
ZHAO MING

Assignees

瑞湖智科数据(苏州)有限公司

Dates

Publication Date: 20260512
Application Date: 20251229

Claims (10)

1. The method for analyzing the bearing state of the wind driven generator is characterized by comprising the following steps of: Acquiring the running parameters of the wind driven generator bearing in real time based on the sensor group to obtain multidimensional running parameter information; preprocessing the multidimensional operation parameter information to obtain preprocessed multidimensional operation parameter information; Respectively extracting time domain features and frequency domain features based on the preprocessed multidimensional operation parameter information to obtain feature sets; performing dimension reduction optimization processing on the feature set by adopting a feature selection algorithm, and removing redundant features and invalid features to obtain an optimal feature subset; Inputting the optimal feature subset into a preset bearing state identification model, and outputting real-time state information of the wind driven generator bearing based on the bearing state identification model; and carrying out state evaluation and abnormal early warning on the real-time state information of the wind driven generator bearing based on a dynamic early warning mechanism to obtain a state analysis result.
2. The method for analyzing the bearing state of the wind driven generator according to claim 1, wherein the sensor group at least comprises a vibration sensor, a temperature sensor and a rotation speed sensor; the operation parameters acquired based on the vibration sensor at least comprise a bearing outer ring vibration signal, an inner ring vibration signal and bearing vibration acceleration; Acquiring a bearing shell temperature signal based on a temperature sensor; Collecting a bearing rotating speed signal based on a rotating speed sensor; and synchronously recording the acquisition time stamp and the operation condition information of the wind driven generator in the acquisition process, wherein the operation condition information comprises wind speed, unit output power and yaw angle, and multi-dimensional operation parameter information is obtained.
3. The method for analyzing the bearing state of the wind driven generator according to claim 2, wherein the preprocessing operation sequentially comprises outlier rejection, data smoothing denoising, signal resampling and data standardization, wherein the outlier rejection adopts a glaubes criterion to identify and reject abnormal data beyond a3 sigma range, and sigma represents a standard deviation; The data smoothing denoising adopts a wavelet threshold denoising algorithm to process vibration signals; The signal resampling unifies parameters of different sampling frequencies to a preset sampling frequency; data normalization parameters were converted to the same order of magnitude using the Z-score normalization method.
4. A method of analyzing the bearing condition of a wind turbine according to claim 3, wherein the time domain features include at least one of peak value, peak factor, kurtosis, skewness, root mean square value and waveform factor; The frequency domain features include at least one of center of gravity frequency, mean square frequency, frequency variance, and peak frequency.
5. The method for analyzing the bearing state of a wind turbine according to claim 4, wherein the feature set is subjected to dimension reduction optimization by using a feature selection algorithm, and the method further comprises the step of verifying feature validity: inputting the optimal feature subset into a verification model; calculating state identification accuracy, recall rate and F1 score corresponding to the feature subset; judging whether the state identification accuracy, recall rate and F1 score are greater than a preset threshold value; If the state identification accuracy, recall and F1 score are all greater than the preset threshold, judging that the feature optimization is effective; If one of the state identification accuracy, recall and F1 score is smaller than or equal to a preset threshold, the feature extraction is carried out again or the feature selection algorithm parameters are adjusted.
6. The method for analyzing the bearing state of the wind driven generator according to claim 5, wherein a bearing state level threshold is set based on a dynamic early warning mechanism, a bearing health evaluation report is generated based on real-time state information and the bearing state level threshold, and an early warning signal is triggered when the real-time state information is greater than or equal to the bearing state level threshold; The bearing health evaluation report comprises a bearing real-time state grade, a real-time numerical value and a change trend curve of an operation parameter, an abnormality degree of a characteristic parameter and a bearing residual service life predicted value, wherein the bearing residual service life predicted value is calculated through a life prediction model constructed based on an optimal characteristic subset, and the life prediction model is an LSTM neural network model or a gray prediction model.
7. The wind driven generator bearing state analysis system is characterized by comprising a memory and a processor, wherein the memory comprises a program of a wind driven generator bearing state analysis method, and the program of the wind driven generator bearing state analysis method is executed by the processor to realize the following steps: Acquiring the running parameters of the wind driven generator bearing in real time based on the sensor group to obtain multidimensional running parameter information; preprocessing the multidimensional operation parameter information to obtain preprocessed multidimensional operation parameter information; Respectively extracting time domain features and frequency domain features based on the preprocessed multidimensional operation parameter information to obtain feature sets; performing dimension reduction optimization processing on the feature set by adopting a feature selection algorithm, and removing redundant features and invalid features to obtain an optimal feature subset; Inputting the optimal feature subset into a preset bearing state identification model, and outputting real-time state information of the wind driven generator bearing based on the bearing state identification model; and carrying out state evaluation and abnormal early warning on the real-time state information of the wind driven generator bearing based on a dynamic early warning mechanism to obtain a state analysis result.
8. The wind turbine bearing condition analysis system of claim 7, wherein the sensor set includes at least a vibration sensor, a temperature sensor, and a rotational speed sensor; the operation parameters acquired based on the vibration sensor at least comprise a bearing outer ring vibration signal, an inner ring vibration signal and bearing vibration acceleration; Acquiring a bearing shell temperature signal based on a temperature sensor; Collecting a bearing rotating speed signal based on a rotating speed sensor; and synchronously recording the acquisition time stamp and the operation condition information of the wind driven generator in the acquisition process, wherein the operation condition information comprises wind speed, unit output power and yaw angle, and multi-dimensional operation parameter information is obtained.
9. The wind turbine bearing state analysis system according to claim 8, wherein the preprocessing operation sequentially comprises outlier rejection, data smoothing denoising, signal resampling and data normalization, wherein outlier rejection uses a glaubes criterion to identify and reject outlier data beyond a3 sigma range, and sigma represents a standard deviation; The data smoothing denoising adopts a wavelet threshold denoising algorithm to process vibration signals; The signal resampling unifies parameters of different sampling frequencies to a preset sampling frequency; data normalization parameters were converted to the same order of magnitude using the Z-score normalization method.
10. A computer readable storage medium, characterized in that it comprises a wind turbine bearing state analysis method program, which, when executed by a processor, implements the steps of a wind turbine bearing state analysis method according to any one of claims 1 to 6.

Description

Wind driven generator bearing state analysis method, system and medium Technical Field The application relates to the technical field of bearing state analysis, in particular to a method, a system and a medium for analyzing the bearing state of a wind driven generator. Background The wind driven generator bearing can undergo stages of running-in, normal operation, failure and the like after being put into operation. The existing CMS system adopts a fixed threshold value or spectrum characteristic comparison mode, and has the following defects: 1) The fixed threshold value can not adapt to individual differences of different models and different bearings; 2) The frequency spectrum comparison only focuses on the fault frequency of parts, so that the integral degradation trend of the bearing is difficult to evaluate; 3) The quantitative division of 'running-in-normal-early fault-late fault' cannot be automatically given, and the on-site predictive maintenance is not facilitated. Disclosure of Invention The embodiment of the application aims to provide a method, a system and a medium for analyzing the bearing state of a wind driven generator, which realize the direct identification of a result by using an artificial intelligence technology through the automatic evaluation and identification of the bearing state of the wind driven generator, reduce the technical requirements of wind field maintenance personnel, realize the high-density equipment state evaluation and improve the reliability of equipment operation and maintenance and evaluation. The embodiment of the application also provides a method for analyzing the bearing state of the wind driven generator, which comprises the following steps: Acquiring the running parameters of the wind driven generator bearing in real time based on the sensor group to obtain multidimensional running parameter information; preprocessing the multidimensional operation parameter information to obtain preprocessed multidimensional operation parameter information; Respectively extracting time domain features and frequency domain features based on the preprocessed multidimensional operation parameter information to obtain feature sets; performing dimension reduction optimization processing on the feature set by adopting a feature selection algorithm, and removing redundant features and invalid features to obtain an optimal feature subset; Inputting the optimal feature subset into a preset bearing state identification model, and outputting real-time state information of the wind driven generator bearing based on the bearing state identification model; and carrying out state evaluation and abnormal early warning on the real-time state information of the wind driven generator bearing based on a dynamic early warning mechanism to obtain a state analysis result. Optionally, in the method for analyzing a bearing state of a wind turbine according to the embodiment of the present application, the sensor group includes at least a vibration sensor, a temperature sensor, and a rotation speed sensor; the operation parameters acquired based on the vibration sensor at least comprise a bearing outer ring vibration signal, an inner ring vibration signal and bearing vibration acceleration; Acquiring a bearing shell temperature signal based on a temperature sensor; Collecting a bearing rotating speed signal based on a rotating speed sensor; and synchronously recording the acquisition time stamp and the operation condition information of the wind driven generator in the acquisition process, wherein the operation condition information comprises wind speed, unit output power and yaw angle, and multi-dimensional operation parameter information is obtained. Optionally, in the method for analyzing the bearing state of the wind turbine according to the embodiment of the present application, the preprocessing operation includes outlier rejection, data smoothing denoising, signal resampling and data normalization in sequence, where outlier rejection uses a gladbis criterion to identify and reject outlier data beyond the 3 sigma range, and sigma represents a standard deviation; The data smoothing denoising adopts a wavelet threshold denoising algorithm to process vibration signals; The signal resampling unifies parameters of different sampling frequencies to a preset sampling frequency; data normalization parameters were converted to the same order of magnitude using the Z-score normalization method. Optionally, in the method for analyzing a bearing state of a wind turbine according to the embodiment of the present application, the time domain feature includes at least one of a peak value, a peak factor, a kurtosis, a skewness, a root mean square value, and a waveform factor; The frequency domain features include at least one of center of gravity frequency, mean square frequency, frequency variance, and peak frequency. Optionally, in the method for analyzing the bearing state of the wind turbine according t