Diagnosis Method and Device for Wind Turbine Drive Train

Abstract

This application proposes a diagnostic method for the drive train of a wind turbine. The method includes: obtaining vibration data at different measurement points of the drive train; extracting vibration characteristics from the vibration data and generating vibration characteristic data according to the vibration characteristics, where the vibration characteristic data includes time-domain characteristics, frequency-domain characteristics, and harmonic characteristics; performing optimal clustering analysis on the vibration characteristic data to obtain the optimal number of clusters, and performing abnormal self-diagnosis according to the optimal number of clusters; if an abnormality is detected, determining the cause of the drive train failure based on the preset analysis method in the vibration expert system. Through abnormal self-diagnosis and fault self-location, this application realizes automatic judgment of fault location, improves the diagnostic analysis efficiency, and at the same time has higher universality, and is applicable to the diagnosis of the drive train of various models.

Description

Technical Field

[0001] This application relates to the technical field of anomaly detection and fault diagnosis, and particularly to a diagnosis method and device for a wind turbine drive train. Background Art

[0002] A wind turbine or a wind power generation unit (hereinafter referred to as the unit) converts wind energy into mechanical energy and then into electrical energy. In this energy conversion process, the drive train plays a crucial role. Any anomaly in the drive train can, at best, lead to the shutdown of the unit and affect its power generation; at worst, it can cause damage to major components, resulting in huge economic losses. Therefore, it is very important to monitor and diagnose the drive train of the unit. Currently, direct-drive units and doubly-fed units generally use a condition monitoring system CMS to monitor and diagnose the vibration of the drive train. The condition monitoring system CMS mainly collects vibration data of the main bearing, gearbox, and generator bearing, and diagnoses the faults of the drive train by analyzing the vibration data.

[0003] In existing related diagnosis methods: some perform anomaly diagnosis of the drive train through empirical threshold methods. For example, referring to the VDI3834 standard, the time-domain characteristics of vibration data are detected, and when the threshold is exceeded, it is determined that there is vibration anomaly in the drive train. Then, a manual or semi-manual method is used for fault location to determine the cause of the drive train anomaly, such as vibration anomaly caused by components such as the main bearing or gearbox. Therefore, the fault location process requires the support of vibration analysis experts, has a certain analysis threshold, and has low analysis efficiency. Some other methods use machine learning methods to combine the anomaly diagnosis and fault location of the drive train into one through model training, replacing experts to complete the fault location process. However, model training is a supervised learning process that requires a rich fault mode library. However, the time cost for accumulating the fault mode library is huge, and the start time of the fault cannot be accurately given. Therefore, the fault mode library is incomplete, and the problem of sample imbalance is extremely obvious. Furthermore, it may lead to overfitting of the model, resulting in a decline in the generalization ability of the model and affecting the universality of the model. Summary of the Invention

[0004] This application aims to solve at least one of the technical problems in the related art to some extent.

[0005] To this end, the first objective of this application is to propose a diagnosis method for a wind turbine drive train, which solves the technical problems of low fault location analysis efficiency and low universality of the fault location model in the existing methods. An adaptive relative threshold method is used for vibration anomaly diagnosis. By comparing the vibration level in the normal state, vibration anomaly diagnosis is realized. At the same time, by regularizing expert experience, automatic judgment of fault location is realized, improving the diagnostic analysis efficiency, and this application has higher universality and is applicable to the diagnosis of drive trains of various models.

[0006] The second object of this application is to propose a diagnostic device for the drive train of a wind turbine generator.

[0007] The third object of this application is to propose a non - transitory computer - readable storage medium.

[0008] To achieve the above object, an embodiment of the first aspect of this application proposes a method for diagnosing the drive train of a wind turbine generator, including: acquiring vibration data of different measurement points on the drive train; extracting vibration features from the vibration data, and generating vibration feature data according to the vibration features, where the vibration feature data includes time - domain features, frequency - domain features, and harmonic features; performing optimal clustering analysis on the vibration feature data to obtain the optimal number of clusters, and performing abnormal self - diagnosis according to the optimal number of clusters; if an abnormality is detected, determining the cause of the drive - train fault based on the preset analysis method in the vibration expert system.

[0009] Optionally, in an embodiment of this application, performing optimal clustering analysis on the vibration feature data to obtain the optimal number of clusters includes:

[0010] Obtaining the optimal clustering result corresponding to the vibration feature data through optimal clustering analysis;

[0011] Counting the number of the optimal clustering result, and taking the number as the optimal number of clusters.

[0012] Optionally, in an embodiment of this application, obtaining the optimal clustering result corresponding to the vibration feature data through optimal clustering analysis includes:

[0013] S1, obtaining the maximum number of loops, the initial value of the number of loops, and the initial value of the number of clusters;

[0014] S2, using a clustering algorithm to perform clustering analysis to obtain a clustering result, and increasing the number of loops by 1;

[0015] S3, calculating an evaluation index according to the clustering result, and saving the evaluation index result;

[0016] S4, determining whether the number of loops is greater than the maximum number of loops;

[0017] S5, if not, increasing the number of clusters by 1, and repeating steps S2 - S4;

[0018] S6, if so, retrieving the optimal index in the evaluation index results, and taking the optimal index as the optimal clustering result.

[0019] Optionally, in an embodiment of this application, performing abnormal self - diagnosis according to the optimal number of clusters includes:

[0020] Obtaining the initial change amount of the vibration feature data;

[0021] Determine whether the optimal number of clusters is greater than a preset threshold;

[0022] If not, it is determined that the vibration of the transmission chain is normal;

[0023] If so, calculate the difference between the maximum cluster center and the minimum cluster center, and compare the difference with the initial change amount;

[0024] If the difference is less than or equal to the initial change amount, it is determined that the vibration of the transmission chain is normal;

[0025] If the difference is greater than the initial change amount, count the frequency of the difference being greater than the initial change amount;

[0026] When the frequency exceeds the preset quantity, it is determined that the vibration of the transmission chain is abnormal.

[0027] Optionally, in an embodiment of the present application, determining the cause of the transmission chain failure based on a preset analysis method in the vibration expert system includes:

[0028] Filter the vibration signal using a specific filter;

[0029] Extract the envelope of the filtered vibration signal and perform FFT transformation to obtain an envelope spectrum;

[0030] Find peaks according to the envelope spectrum to obtain multiple groups of peaks;

[0031] Match the frequencies corresponding to the multiple groups of peaks with the fault frequency library to achieve fault location.

[0032] Optionally, in an embodiment of the present application, determining the cause of the transmission chain failure based on a preset analysis method in the vibration expert system includes:

[0033] Perform FFT transformation on the vibration signal, extract the frequency values in a specific frequency band to obtain the spectrum in the specific frequency band;

[0034] Find peaks in the spectrum in the specific frequency band to obtain multiple groups of peaks;

[0035] Match the frequencies corresponding to the multiple groups of peaks with the fault frequency library to achieve fault location.

[0036] Optionally, in an embodiment of the present application, determining the cause of the transmission chain failure based on a preset analysis method in the vibration expert system includes:

[0037] Perform FFT transformation on the vibration signal, calculate the meshing frequency of the gearbox, and extract the left and right sidebands at the meshing frequency to obtain the meshing frequency sidebands;

[0038] Find peaks in the meshing frequency sidebands to obtain multiple groups of peaks;

[0039] Match the frequencies corresponding to multiple groups of peaks with the fault frequency library to achieve fault location.

[0040] To achieve the above object, an embodiment of the second aspect of the present application provides a diagnosis device for a wind turbine drive train, including:

[0041] An acquisition module, configured to acquire vibration data of different measurement points of the drive train;

[0042] A feature extraction module, configured to extract vibration features from the vibration data and generate vibration feature data according to the vibration features, where the vibration feature data includes time-domain features, frequency-domain features, and harmonic features;

[0043] An abnormal self-diagnosis module, configured to perform optimal clustering analysis on the vibration feature data to obtain the optimal clustering number, and perform abnormal self-diagnosis according to the optimal clustering number;

[0044] A fault self-location module, configured to, if an abnormality is detected, determine the cause of the drive train fault based on a preset analysis method in the vibration expert system.

[0045] Optionally, in an embodiment of the present application, the abnormal self-diagnosis module is specifically configured to:

[0046] Obtain the optimal clustering result corresponding to the vibration feature data through optimal clustering analysis;

[0047] Count the number of the optimal clustering results, and use the number as the optimal clustering number.

[0048] To achieve the above object, an embodiment of the third aspect of the present application provides a non-transitory computer-readable storage medium, which, when the instructions in the storage medium are executed by a processor, enables the execution of a diagnosis method for a wind turbine drive train.

[0049] The diagnosis method for a wind turbine drive train, the diagnosis device for a wind turbine drive train, and the non-transitory computer-readable storage medium according to the embodiments of the present application solve the technical problems of low efficiency of fault location analysis and low universality of the fault location model in the existing methods. An adaptive relative threshold method is used for vibration abnormality diagnosis. By comparing the vibration level in the normal state, vibration abnormality diagnosis is achieved. At the same time, by regularizing expert experience, automatic judgment of fault location is realized, the diagnosis and analysis efficiency is improved, and the present application has higher universality and is applicable to the diagnosis of drive trains of various models.

[0050] Additional aspects and advantages of the present application will be given in part in the following description, become apparent in part from the following description, or be learned through the practice of the present application. Description of the Drawings

[0051] The above and / or additional aspects and advantages of the present application will become apparent and be readily understood from the following description of embodiments in conjunction with the accompanying drawings, where:

[0052] Figure 1 It is a flowchart of a wind turbine drive train diagnosis method provided in the first embodiment of the present application;

[0053] Figure 2 It is an example diagram of abnormal self-diagnosis of the wind turbine drive train diagnosis method in the embodiment of the present application;

[0054] Figure 3 It is an example diagram of the change in vibration level under normal and abnormal states of the wind turbine drive train diagnosis method in the embodiment of the present application;

[0055] Figure 4 It is an example diagram of optimal clustering analysis of the wind turbine drive train diagnosis method in the embodiment of the present application;

[0056] Figure 5 It is an example diagram of the comprehensive diagnosis process of the wind turbine drive train diagnosis method in the embodiment of the present application;

[0057] Figure 6 It is a schematic structural diagram of a wind turbine drive train diagnosis device provided in the second embodiment of the present application. Detailed Embodiments

[0058] The embodiments of the present application will be described in detail below. The examples of the embodiments are shown in the accompanying drawings, where 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 are intended to explain the present application and should not be construed as limiting the present application.

[0059] The wind turbine drive train diagnosis method and device of the embodiments of the present application will be described below with reference to the accompanying drawings.

[0060] Figure 1 It is a schematic flowchart of a wind turbine drive train diagnosis method provided in the first embodiment of the present application.

[0061] As Figure 1 shown, the wind turbine drive train diagnosis method includes the following steps:

[0062] Step 101, obtain vibration data of different measurement points of the drive train.

[0063] In the embodiments of the present application, the vibration data is the vibration data of different measuring points of the wind turbine drive train obtained from the CMS database. To obtain the vibration data of different measuring points of the drive train, the specific operation is to collect the vibration data of the drive train by installing multiple sensors in different directions such as axial and radial, taking into account the main bearing, gearbox, and generator bearing. After that, data cleaning is performed on the vibration data of different measuring points, including shutdown data cleaning, out-of-range data elimination, and noise data cleaning.

[0064] Step 102: Extract vibration features from the vibration data and generate vibration feature data according to the vibration features. The vibration feature data includes time-domain features, frequency-domain features, and harmonic features.

[0065] In the embodiments of the present application, in combination with the fault mechanisms of bearings and gearboxes, multi-sensor feature extraction is performed. Specifically, different features are extracted from multiple measuring points such as the main bearing, gearbox, and generator. The extracted features include filtered time-domain features, frequency-domain features within a specific frequency band, harmonic features at the meshing frequency, etc.

[0066] Step 103: Perform optimal clustering analysis on the vibration feature data to obtain the optimal clustering number, and perform abnormal self-diagnosis according to the optimal clustering number.

[0067] In the embodiments of the present application, abnormal self-diagnosis is performed according to the vibration feature data. Among them, considering that the vibration levels of different units are different, different relative thresholds are given, and vibration abnormal diagnosis can be performed adaptively.

[0068] Exemplarily, the present application can select the vibration feature data for a period of time, use the vibration feature data as the input, and perform optimal clustering analysis. When the drive train is abnormal, the vibration level will increase. Among the vibration feature data selected for a period of time, the vibration level of the faulty data will inevitably increase. Through optimal clustering analysis, the vibration level analysis is performed on the vibration data for this period of time to obtain the optimal clustering result, and the number of the optimal clustering results is counted. The number is used as the optimal clustering number. If the optimal clustering number is greater than 2, it indicates that the vibration level has shown significant differences during this period of time, and the vibration of the data with a larger clustering center has increased significantly.

[0069] Step 104: If an abnormality is detected, determine the cause of the drive train fault based on the analysis method preset in the vibration expert system.

[0070] In the embodiments of the present application, after detecting vibration abnormalities through optimal clustering analysis, fault self-localization is performed. Among them, the self-localization method, based on vibration expert experience, regularizes the process of manual analysis, constructs a vibration expert system using the analysis method transformed by rules, and then realizes the fault self-localization process. And before regularization, full communication with vibration experts is required to determine the effective analysis methods in vibration analysis and perform rule transformation.

[0071] The wind turbine drive train diagnosis method according to the embodiments of the present application obtains vibration data of different measuring points on the drive train; extracts vibration characteristics from the vibration data, and generates vibration characteristic data according to the vibration characteristics, wherein the vibration characteristic data includes time domain characteristics, frequency domain characteristics, and harmonic characteristics; performs optimal clustering analysis on the vibration characteristic data to obtain the optimal clustering number, and performs abnormal self-diagnosis according to the optimal clustering number; if an abnormality is detected, determines the cause of the drive train fault based on the preset analysis method in the vibration expert system. Thus, the technical problems of low fault location analysis efficiency and low universality of the fault location model in the existing methods can be solved. The adaptive relative threshold method is used for vibration abnormality diagnosis. By comparing the vibration level in the normal state, vibration abnormality diagnosis is realized. At the same time, by regularizing expert experience, automatic fault location judgment is realized, the diagnosis and analysis efficiency is improved, and the present application has higher universality and is applicable to the diagnosis of drive trains of various models.

[0072] The wind turbine drive train diagnosis method provided by the present application includes an abnormal self-diagnosis part and a fault self-location part. Among them, the abnormal self-diagnosis part uses the adaptive relative threshold method for vibration abnormality diagnosis, and realizes vibration abnormality diagnosis by comparing the vibration level in the normal state. Further, the adaptive relative threshold abnormality diagnosis method uses different relative thresholds according to the vibration levels of different units, and is an adaptive diagnosis method. The fault location part combines expert experience, regularizes the experience, realizes automatic fault location, improves the diagnosis efficiency, and at the same time, through continuous experience accumulation and rule testing, the universality of the diagnosis rules will be continuously improved. More importantly, the wind turbine drive train diagnosis method provided by the present application is applicable to various units such as direct drive units and doubly fed units.

[0073] Further, in the embodiments of the present application, performing optimal clustering analysis on the vibration characteristic data to obtain the optimal clustering number includes:

[0074] Obtaining the optimal clustering result corresponding to the vibration characteristic data through optimal clustering analysis;

[0075] Counting the number of the optimal clustering results, and taking the number as the optimal clustering number.

[0076] It can be understood that in this application, vibration characteristic data for a period of time can be selected, and the vibration characteristic data can be used as input for optimal clustering analysis. When the transmission chain is abnormal, the vibration level will increase. Among the vibration characteristic data for the selected period of time, the vibration level of the faulty data will inevitably increase. Through optimal clustering analysis, the vibration level of the vibration data for this period is analyzed to obtain the optimal clustering result, and the number of the optimal clustering result is counted, and the number is used as the optimal clustering number. If the optimal clustering number is greater than 2, it indicates that the vibration level for this period has shown significant differences, and the vibration of the data with a larger clustering center has increased significantly.

[0077] Further, in the embodiment of this application, the optimal clustering result corresponding to the vibration characteristic data is obtained through optimal clustering analysis, including:

[0078] S1, obtaining the maximum number of cycles, the initial value of the number of cycles, and the initial value of the number of clusters;

[0079] S2, using a clustering algorithm to perform clustering analysis to obtain a clustering result, and increasing the number of cycles by 1;

[0080] S3, calculating an evaluation index according to the clustering result and saving the evaluation index result;

[0081] S4, determining whether the number of cycles is greater than the maximum number of cycles;

[0082] S5, if not, increasing the number of clusters by 1, and repeating steps S2 - S4;

[0083] S6, if so, retrieving the optimal index in the evaluation index results and using the optimal index as the optimal clustering result.

[0084] In the embodiment of this application, a clustering algorithm is selected for clustering analysis, and the selected clustering algorithm can be a mixture of Gaussian clustering algorithm, Kmeans clustering algorithm, etc. The evaluation index is calculated and the evaluation index result is saved. The selected evaluation index can be the AIC criterion (Akaike information criterion), the BIC criterion (Bayesian information criterion), etc.

[0085] Exemplarily, the optimal clustering result corresponding to the vibration characteristic data obtained through optimal clustering analysis can be that, given the maximum number of cycles and the initial value of the number of cycles, the number of clusters is given within each cycle, and the number of clusters gradually increases starting from 1. A clustering algorithm is selected for clustering analysis. After obtaining the clustering result, the evaluation index is calculated and the evaluation index result is saved. Then it is determined whether the number of cycles is greater than the maximum number of cycles. If not, continue to perform clustering analysis in a loop; if it is greater than the maximum number of cycles, find the optimal index among the evaluation indexes obtained in all cycles, and the clustering result corresponding to the optimal index is the optimal clustering result.

[0086] Further, in the embodiments of the present application, anomaly self-diagnosis based on the optimal clustering number includes:

[0087] Obtain the initial change amount of the vibration characteristic data;

[0088] Determine whether the optimal clustering number is greater than a preset threshold;

[0089] If not, it is determined that the vibration of the transmission chain is normal;

[0090] If so, calculate the difference between the maximum clustering center and the minimum clustering center, and compare the difference with the initial change amount;

[0091] If the difference is less than or equal to the initial change amount, it is determined that the vibration of the transmission chain is normal;

[0092] If the difference is greater than the initial change amount, count the frequency of the difference being greater than the initial change amount;

[0093] When the frequency exceeds a preset number, it is determined that the vibration of the transmission chain is abnormal.

[0094] Exemplarily, anomaly self-diagnosis based on the optimal clustering number can be to first determine whether the optimal clustering number is greater than 2. If it is greater than 2, it means that the vibration level has changed significantly during this period, and it is suspected that the transmission chain has abnormal vibration. Otherwise, it is determined that the given vibration level is normal. When the optimal clustering number is greater than 2, calculate the difference between the minimum clustering center and the maximum clustering center. If the difference is greater than the initial change amount, continue to count the frequency exceeding the initial change amount. When the frequency exceeds a certain number, it is determined that the transmission chain has an abnormality.

[0095] In the embodiments of the present application, after detecting vibration anomalies through optimal clustering analysis, fault self-localization is performed. Among them, the self-localization method, based on vibration expert experience, regularizes the process of manual analysis, constructs a vibration expert system using the analysis method transformed by rules, and then realizes the fault self-localization process. However, some expert experiences are relatively complex, and there are conflicts in the descriptions of different experts. Then, the quantifiable implementation methods can be regularized for automatic diagnosis. And before regularization, sufficient communication with vibration experts is required to determine the effective analysis methods in vibration analysis and perform rule transformation.

[0096] Further, in the embodiments of the present application, determining the cause of the transmission chain fault based on the preset analysis method in the vibration expert system includes:

[0097] Filter the vibration signal using a specific filter;

[0098] Extract the envelope of the filtered vibration signal and perform FFT transformation to obtain the envelope spectrum;

[0099] Find peaks according to the envelope spectrum to obtain multiple groups of peaks;

[0100] The frequencies corresponding to multiple groups of peaks are matched with the fault frequency library to achieve fault location.

[0101] One method for determining the cause of a transmission chain failure based on the analysis method preset in the vibration expert system is to first perform filtering with a specific filter, such as a low-pass filter, and then perform envelope extraction after filtering. Then, perform an FFT transform to obtain an envelope spectrum. Peak search is performed on the envelope spectrum to obtain multiple groups of peaks. Finally, the frequencies corresponding to the multiple groups of peaks are matched with the fault frequency library to achieve fault location. For example, when a bearing inner ring fails, the frequency of the inner ring failure and its harmonics will be present in the envelope spectrum. These harmonics are often the multiple peaks in the envelope spectrum. Matching through peak search can achieve a self-locating fault process.

[0102] Furthermore, in the embodiment of the present application, determining the cause of the transmission chain failure based on the analysis method preset in the vibration expert system includes:

[0103] Perform FFT transformation on the vibration signal, extract the frequency value within a specific frequency band, and obtain the spectrum within the specific frequency band;

[0104] Perform peak search on the spectrum within a specific frequency band to obtain multiple groups of peaks;

[0105] The frequencies corresponding to multiple groups of peaks are matched with the fault frequency library to achieve fault location.

[0106] One method for determining the cause of a transmission chain fault based on the analysis methods pre-defined in the vibration expert system involves performing an FFT transform on the vibration signal to extract frequency values within a specific frequency band. Peak search of the spectrum within this frequency band is then performed, and the peaks are matched against a fault frequency library to achieve self-localization of the fault. For example, if a tooth on a gearbox's high-speed shaft pinion breaks, the high-speed shaft's rotational frequency and its harmonic components will appear within a specific flat band. By extracting the spectrum within this specific frequency band, performing a peak search, and matching it against a fault library, the gear problem can be diagnosed.

[0107] Furthermore, in the embodiment of the present application, determining the cause of the transmission chain failure based on the analysis method preset in the vibration expert system includes:

[0108] Perform FFT transformation on the vibration signal to calculate the meshing frequency of the gearbox, and extract the left and right sidebands at the meshing frequency to obtain the meshing frequency sidebands;

[0109] Perform peak search on the meshing frequency sidebands to obtain multiple groups of peaks;

[0110] The frequencies corresponding to multiple groups of peaks are matched with the fault frequency library to achieve fault location.

[0111] An implementation method for determining the cause of transmission chain faults based on the preset analysis method in the vibration expert system can be as follows: perform FFT transformation on the vibration signal, calculate the meshing frequency of the gearbox, and extract the left and right sidebands at the meshing frequency. Then, perform peak searching and match with the fault frequency library. Finally, fault location can be achieved. Exemplarily, if there is a fault in a certain bearing in the gearbox, sidebands will be formed at the meshing of the gearbox. Extract these sidebands for peak searching and match with the fault frequency library to achieve the fault location of the gearbox bearing.

[0112] Figure 2 This is an example diagram of abnormal self-diagnosis for the wind turbine drive train diagnosis method according to the embodiments of the present application.

[0113] As Figure 2 shown, obtain vibration characteristic data, perform optimal clustering analysis on the vibration characteristic data to obtain the optimal number of clusters, and determine whether the optimal number of clusters is greater than 2. If it is greater than 2, it indicates that there is a significant difference in the vibration level during this period, suspecting that the drive train has vibration abnormalities. Otherwise, give a judgment that the vibration level is normal. When the optimal number of clusters is greater than 2, calculate the difference between the minimum cluster center and the maximum cluster center. If the difference is greater than the initial change amount, continue to count the frequency exceeding the initial change amount. When the frequency exceeds a certain value, it is judged that the drive train is abnormal. If the difference is less than or equal to the initial change amount, give a judgment that the vibration level is normal.

[0114] Figure 3 This is an example diagram of the change in vibration level under normal and abnormal states for the wind turbine drive train diagnosis method according to the embodiments of the present application.

[0115] As Figure 3 shown, when performing optimal clustering analysis in the healthy state, only one clustering result can be obtained; while when performing optimal clustering analysis in the fault state, two or more clustering results will be obtained.

[0116] Figure 4 This is an example diagram of optimal clustering analysis for the wind turbine drive train diagnosis method according to the embodiments of the present application.

[0117] As Figure 4 shown, given the maximum number of cycles and the initial value of the number of cycles, the number of clusters is given within each cycle, and the number of clusters gradually increases starting from 1. Select a clustering algorithm for clustering analysis. After obtaining the clustering result, calculate the evaluation index and save the evaluation index result. Then determine whether the number of cycles is greater than the maximum number of cycles. If not, continue to perform clustering analysis in a loop; if it is greater than the maximum number of cycles, find the optimal index among all the evaluation indexes obtained in the loops. The clustering result corresponding to the optimal index is the optimal clustering result.

[0118] Figure 5This is an example diagram of the comprehensive diagnosis process for the wind turbine drive train diagnosis method according to the embodiments of the present application.

[0119] As Figure 5 shown, based on the experience of vibration experts, the process of manual analysis is regularized. Among them, the regularization process includes filtering the vibration signal using a specific filter, performing envelope extraction on the filtered vibration signal, and performing FFT transformation to obtain an envelope spectrum. Peak search is performed according to the envelope spectrum to obtain multiple groups of peaks. The frequencies corresponding to the multiple groups of peaks are matched with the fault frequency library to achieve fault location; FFT transformation is performed on the vibration signal, the frequency values within a specific frequency band are extracted to obtain the spectrum within the specific frequency band, peak search is performed on the spectrum within the specific frequency band to obtain multiple groups of peaks, and the frequencies corresponding to the multiple groups of peaks are matched with the fault frequency library to achieve fault location; FFT transformation is performed on the vibration signal, the meshing frequency of the gearbox is calculated, and the left and right sidebands at the meshing frequency are extracted to obtain the meshing frequency sidebands. Peak search is performed on the meshing frequency sidebands to obtain multiple groups of peaks, and the frequencies corresponding to the multiple groups of peaks are matched with the fault frequency library to achieve fault location.

[0120] Figure 6 This is a schematic structural diagram of a wind turbine drive train diagnosis device provided in the second embodiment of the present application.

[0121] As Figure 6 shown, the wind turbine drive train diagnosis device includes:

[0122] An acquisition module 10 for acquiring vibration data of different measuring points of the drive train;

[0123] A feature extraction module 20 for extracting vibration features from the vibration data and generating vibration feature data according to the vibration features. Among them, the vibration feature data includes time-domain features, frequency-domain features, and harmonic features;

[0124] An abnormal self-diagnosis module 30 for performing optimal clustering analysis on the vibration feature data to obtain the optimal clustering number and performing abnormal self-diagnosis according to the optimal clustering number;

[0125] A fault self-location module 40 for, if an abnormality is detected, determining the cause of the drive train fault based on the preset analysis method in the vibration expert system.

[0126] The wind turbine drive train diagnosis device according to the embodiment of the present application includes an acquisition module for acquiring vibration data of different measurement points on the drive train; a feature extraction module for extracting vibration features from the vibration data and generating vibration feature data according to the vibration features, wherein the vibration feature data includes time domain features, frequency domain features, and harmonic features; an abnormal self-diagnosis module for performing optimal clustering analysis on the vibration feature data to obtain the optimal clustering number and performing abnormal self-diagnosis according to the optimal clustering number; and a fault self-localization module for, if an abnormality is detected, determining the cause of the drive train fault based on the preset analysis method in the vibration expert system. Thus, the technical problems of low efficiency of fault location analysis and low universality of the fault location model in the existing methods can be solved. The adaptive relative threshold method is used for vibration abnormality diagnosis. By comparing the vibration level in the normal state, vibration abnormality diagnosis is realized. At the same time, by regularizing expert experience, automatic judgment of fault location is realized, the diagnosis and analysis efficiency is improved, and the present application has higher universality and is applicable to the diagnosis of drive trains of various models.

[0127] Further, in the embodiment of the present application, the abnormal self-diagnosis module is specifically used for:

[0128] Obtaining the optimal clustering result corresponding to the vibration feature data through optimal clustering analysis;

[0129] Counting the number of the optimal clustering results and taking the number as the optimal clustering number.

[0130] To implement the above embodiment, the present application also proposes a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the wind turbine drive train diagnosis method of the above embodiment is implemented.

[0131] In the description of this specification, the descriptions referring to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any one or more embodiments or examples in a suitable manner. In addition, without conflict, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0132] In addition, 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 quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one such feature. In the description of this application, "a plurality of" means at least two, such as two, three, etc., unless otherwise specifically defined.

[0133] Any process or method description represented in a flowchart or described otherwise herein may be understood to represent code for an executable instruction including one or more steps for implementing a customized logic function or process, a module, segment, or portion of code, and the scope of the preferred embodiments of this application includes additional implementations, where functions may be executed in a substantially simultaneous manner or in a reverse order according to the functions involved, rather than in the order shown or discussed, which should be understood by those skilled in the art to which the embodiments of this application pertain.

[0134] Logic and / or steps represented in a flowchart or described otherwise herein, for example, may be considered a sequenced list of executable instructions for implementing a logical function, and may be embodied specifically in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium include the following: an electrical connection portion having one or more wirings (electronic device), a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium can even be paper or other suitable medium on which the program can be printed, as the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpretation, or other appropriate processing as necessary, and then stored in a computer memory.

[0135] It should be understood that each part of the present application can be implemented by hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one of the following techniques known in the art or a combination thereof can be used: discrete logic circuits with logic gate circuits for implementing logic functions on data signals, application-specific integrated circuits with appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0136] Those of ordinary skill in the art can understand that all or part of the steps carried by the method of the above embodiments can be completed by instructing relevant hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiments.

[0137] In addition, in each embodiment of the present application, each functional unit can be integrated into a processing module, or each unit can exist physically alone, or two or more units can be integrated into one module. The above integrated module can be implemented in the form of hardware or in the form of a software functional module. When the above integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0138] The above-mentioned storage medium can be a read-only memory, a magnetic disk, an optical disc, etc. Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and should not be construed as limiting the present application. Those of ordinary skill in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of the present application.

Claims

1. A diagnosis method for the drive train of a wind turbine, characterized in that, It includes the following steps: Obtain the vibration data of different measuring points on the transmission chain; Extract vibration characteristics from the vibration data, and generate vibration characteristic data according to the vibration characteristics. Among them, the vibration characteristic data includes time-domain characteristics, frequency-domain characteristics, and harmonic characteristics; Perform optimal clustering analysis on the vibration characteristic data to obtain an optimal clustering result, count the number of the optimal clustering result, use the number as the optimal clustering number, and perform abnormal self-diagnosis according to the optimal clustering number; If an abnormality is detected, determine the cause of the transmission chain failure based on the preset analysis method in the vibration expert system; among them, perform FFT transformation on the vibration signal, extract the frequency values within a specific frequency band to obtain the spectrum within the specific frequency band; perform peak search on the spectrum within the specific frequency band to obtain multiple groups of peaks; match the frequencies corresponding to the multiple groups of peaks with the fault frequency library to achieve fault location; Perform optimal clustering analysis on the vibration characteristic data to obtain an optimal clustering result, including: S1. Obtain the maximum number of cycles, the initial value of the number of cycles, and the initial value of the number of clusters; S2. Use the clustering algorithm to perform clustering analysis to obtain a clustering result, and increase the number of cycles by 1; S3. Calculate the evaluation index according to the clustering result, and save the evaluation index result; S4. Determine whether the number of cycles is greater than the maximum number of cycles; S5. If not, increase the number of clusters by 1, and repeat steps S2 - S4; S6. If so, retrieve the optimal index from the evaluation index results, and use the optimal index as the optimal clustering result; Perform abnormal self-diagnosis according to the optimal clustering number, including: Obtain the initial change amount of the vibration characteristic data; Determine whether the optimal clustering number is greater than the preset threshold; If not, determine that the vibration of the transmission chain is normal; If so, calculate the difference between the maximum clustering center and the minimum clustering center, and compare the difference with the initial change amount; If the difference is less than or equal to the initial change amount, determine that the vibration of the transmission chain is normal; If the difference is greater than the initial change amount, count the frequency of the difference being greater than the initial change amount; When the frequency exceeds the preset number, determine that the vibration of the transmission chain is abnormal.

2. The method according to claim 1, characterized in that, The determination of the cause of the transmission chain failure based on the preset analysis method in the vibration expert system further includes: Filter the vibration signal using a specific filter; Perform envelope extraction on the filtered vibration signal, and perform FFT transformation to obtain an envelope spectrum; Perform peak search according to the envelope spectrum to obtain multiple groups of peaks; Match the frequencies corresponding to the multiple groups of peaks with the fault frequency library to achieve fault location.

3. The method according to claim 1, wherein The determination of the cause of the transmission chain failure based on the preset analysis method in the vibration expert system includes: Perform FFT transformation on the vibration signal, calculate the meshing frequency of the gearbox, and extract the left and right sidebands at the meshing frequency to obtain the meshing frequency sidebands; Perform peak search on the meshing frequency sidebands to obtain multiple groups of peaks; Match the frequencies corresponding to the multiple groups of peaks with the fault frequency library to achieve fault location.

4. A diagnostic device for a wind turbine drive train, characterized in that, It includes: An acquisition module for acquiring the vibration data of different measuring points on the transmission chain; A feature extraction module, configured to extract vibration features from the vibration data and generate vibration feature data according to the vibration features, wherein the vibration feature data includes time-domain features, frequency-domain features, and harmonic features; An abnormal self-diagnosis module, configured to perform optimal clustering analysis on the vibration feature data to obtain an optimal clustering result, count the number of the optimal clustering results, use the number as the optimal clustering number, and perform abnormal self-diagnosis according to the optimal clustering number; A fault self-localization module, configured to, if an abnormality is detected, determine the cause of the transmission chain fault based on a preset analysis method in a vibration expert system; wherein, perform an FFT transformation on the vibration signal, extract frequency values within a specific frequency band to obtain a spectrum within the specific frequency band; perform peak searching on the spectrum within the specific frequency band to obtain multiple groups of peaks; match the frequencies corresponding to the multiple groups of peaks with a fault frequency library to achieve fault localization; Performing optimal clustering analysis on the vibration feature data to obtain an optimal clustering result, including: S1, obtaining a maximum number of loops, an initial value of the number of loops, and an initial value of the number of clusters; S2, using a clustering algorithm to perform clustering analysis to obtain a clustering result, and increasing the number of loops by 1; S3, calculating an evaluation index according to the clustering result and saving the evaluation index result; S4, determining whether the number of loops is greater than the maximum number of loops; S5, if not, increasing the number of clusters by 1, and repeating steps S2-S4; S6, if so, retrieving an optimal index from the evaluation index results, and using the optimal index as the optimal clustering result; Performing abnormal self-diagnosis according to the optimal clustering number, including: obtaining an initial change amount of the vibration feature data; determining whether the optimal clustering number is greater than a preset threshold; if not, determining that the vibration of the transmission chain is normal; if so, calculating a difference between a maximum cluster center and a minimum cluster center, and comparing the difference with the initial change amount; if the difference is less than or equal to the initial change amount, determining that the vibration of the transmission chain is normal; if the difference is greater than the initial change amount, counting the frequency of the difference being greater than the initial change amount; when the frequency exceeds a preset number, determining that the vibration of the transmission chain is abnormal.

5. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, the method as described in any one of claims 1-3 is implemented.

Images