Fan blade voiceprint defect recognition method and system
By establishing a defect identification database and system based on historical voiceprint data, the problems of insufficient real-time performance and accuracy in wind turbine blade defect monitoring have been solved, achieving efficient and reliable defect identification and real-time monitoring.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUADIAN SHANDONG NEW ENERGY CO LTD FEICHENG BRANCH
- Filing Date
- 2026-01-28
- Publication Date
- 2026-05-08
AI Technical Summary
Existing wind turbine blade defect monitoring methods suffer from poor real-time performance and low accuracy. In particular, defect identification relies on manual judgment and is highly complex in intelligent data acquisition and analysis.
By acquiring historical voiceprint recognition data to classify defect types and extract features, a comprehensive defect recognition database is established. Real-time monitoring is performed using big data, and a tight voiceprint defect recognition system is formed by combining data acquisition, classification, and feature extraction units.
It improves the accuracy of defect identification and real-time monitoring, has strong adaptability and fast and efficient identification capabilities, reduces manual intervention, and improves analysis efficiency.
Smart Images
Figure CN121993364A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of wind turbine acoustic defect identification technology, and more specifically, to a method and system for identifying acoustic defects in wind turbine blades. Background Technology
[0002] Wind turbine blades are the most important structural components of a wind turbine. Because they are constantly in motion and subjected to wind force, they are prone to failure, and maintenance after a failure is extremely inconvenient. Therefore, proactive monitoring and maintenance are necessary to improve wind turbine efficiency and reduce operating costs. Of course, wind turbine blades will exhibit operational defects before failure; real-time monitoring of these defects allows for effective proactive maintenance.
[0003] Currently, defect monitoring of wind turbine blades mainly includes manual inspection and intelligent data acquisition and analysis. Intelligent data acquisition and analysis primarily involves acquiring acoustic fingerprint data of wind turbine blades and performing a series of processes, including noise reduction and modal analysis, to determine whether defects exist based on the processed data. Although this method saves labor costs compared to manual inspection, it lacks real-time performance due to the complexity of the analysis methods, and the need for manual judgment of the analysis results makes the accuracy of defect identification unreliable.
[0004] Therefore, designing a method and system for identifying acoustic defects in wind turbine blades, and using more intelligent and efficient acoustic data processing and analysis methods to improve analysis efficiency while ensuring the reliability and accuracy of defect identification, is an urgent problem to be solved. Summary of the Invention
[0005] The purpose of this invention is to provide a method for identifying acoustic fingerprint defects in wind turbine blades. This method involves acquiring historical acoustic fingerprint data to segment the data for different defect types and extracting corresponding defect features, thereby establishing a comprehensive defect identification database using big data. This database is then used to monitor the presence of defects in the data requiring defect identification in real time. On one hand, big data allows for accurate acquisition of characteristics of different types of defects, providing reliable and accurate comparative data for defect identification and ensuring its accuracy. On the other hand, compared to traditional methods such as modal analysis, this method is more adaptable and can quickly and efficiently identify defects, significantly improving the real-time monitoring effect.
[0006] The present invention also aims to provide a wind turbine blade acoustic print defect identification system. This system combines different functional units to form a tightly integrated system capable of identifying defects in acoustic print data. The data acquisition unit continuously acquires historical data to expand the basic big data; the classification and segmentation unit cleans the basic big data; the feature extraction unit extracts targeted features from the segmented big data; and the defect identification unit quickly and accurately identifies defects in the target acoustic print data, greatly improving analysis efficiency while ensuring the reliability and accuracy of defect identification.
[0007] In a first aspect, the present invention provides a method for identifying acoustic defects in wind turbine blades, comprising: collecting historical acoustic identification data, classifying and dividing defect types to form historical acoustic identification classification data; extracting features for defect types based on historical acoustic identification classification data to form acoustic defect identification feature data; acquiring target acoustic data, and combining it with acoustic defect identification feature data to perform defect identification analysis to form target acoustic defect identification data.
[0008] In this invention, the method acquires historical voiceprint recognition data to segment voiceprint data for different defect types and extract corresponding defect features, establishing a comprehensive defect identification database using big data. This database is then used to monitor the data requiring defect identification in real time to determine the presence of defects. On one hand, big data allows for accurate acquisition of characteristics of different types of defects, providing reliable and accurate comparative data for defect identification and ensuring its accuracy. On the other hand, compared to traditional methods such as modal analysis, this approach is more adaptable and enables rapid and efficient defect identification, significantly improving the real-time monitoring effect of defect identification.
[0009] One possible approach is to collect historical voiceprint recognition data, classify it based on defect type, and form historical voiceprint recognition classification data. This includes: extracting voiceprint data with normal monitoring results from historical voiceprint recognition data to form historical normal voiceprint data; extracting voiceprint data of different defect types from historical voiceprint recognition data to form corresponding historical defect type voiceprint data; and combining historical normal voiceprint data and historical defect type voiceprint data of different defect types to form historical voiceprint recognition classification data.
[0010] In this invention, the classification of historical data mainly involves determining whether the voiceprint data has defects. The classification considers two aspects: firstly, extracting normal voiceprint data from historical data to extract features of normal voiceprint data, providing a positive reference for subsequent defect identification; secondly, extracting voiceprint data of different defect types to provide more targeted basic big data for subsequent feature extraction, ensuring the accuracy and rationality of voiceprint feature data for different defect types.
[0011] As one possible approach, based on historical voiceprint recognition classification data, feature extraction is performed for defect types to form voiceprint defect recognition feature data. This includes: performing feature analysis on blade state based on historical normal voiceprint data to form normal voiceprint feature data; combining normal voiceprint feature data to perform defect feature analysis on voiceprint data of different historical defect types to form defect voiceprint feature data corresponding to different defect types; and combining normal voiceprint feature data and different defect voiceprint feature data to form voiceprint defect recognition feature data.
[0012] In this invention, the extraction of voiceprint features includes two aspects. One is the extraction of normal voiceprint features, which can be obtained directly from historical normal voiceprint big data. The other is the extraction of defect voiceprint feature data for different defect types. Since the voiceprint features exhibited by different defect types are different, and defects are only a small part of the entire voiceprint information, most of the voiceprint data is still normal information. Therefore, when performing analysis, it is necessary to consider combining normal voiceprint feature data to assist in the extraction of defect voiceprint features.
[0013] As one possible approach, based on historical normal acoustic fingerprint data, feature analysis targeting the blade state is performed to form normal acoustic fingerprint feature data. This includes: obtaining the normal time-domain variation function of the acoustic fingerprint for each consecutive time period based on historical normal acoustic fingerprint data; obtaining the normal state parameter variation function of different blade state parameters for different normal time-domain variation functions of the acoustic fingerprint; establishing the corresponding time-period acoustic fingerprint-normal state relationship based on the corresponding normal state parameter variation function for different normal time-domain variation functions of the acoustic fingerprint; and determining the acoustic fingerprint-normal state relationship based on the different time-period acoustic fingerprint-normal state relationship.
[0014] In this invention, for normal acoustic fingerprint data, the main factors affecting its acoustic fingerprint characteristics are the blade's operating state, such as rotational speed, acceleration, wind resistance, temperature, and humidity. Therefore, to obtain the characteristic information of normal acoustic fingerprints, the influence of these state parameters on the acoustic fingerprint characteristics can be fully considered. First, normal acoustic fingerprint data for each consecutive time period is acquired to determine the correlation between different state parameters and the temporal variation function of the acoustic fingerprint in the time dimension under multiple sample data items. Considering that although the acquired normal acoustic fingerprint data has undergone noise reduction processing, it may still be affected by noise, and may also be affected by other factors, resulting in certain differences in the correlation relationship obtained in different time periods, the correlation characteristic relationship of normal acoustic fingerprints is determined by obtaining an average function. Although the correlation relationship established in this way is somewhat lacking in accuracy in expressing the relationship between acoustic fingerprint information and state parameters, it can still serve as important reference data for eliminating interference from normal acoustic fingerprint information during subsequent defect feature extraction.
[0015] As one possible approach, by combining normal voiceprint feature data, defect feature analysis is performed on voiceprint data of different historical defect types to form defect voiceprint feature data corresponding to different defect types. This includes: obtaining the time-domain variation function of voiceprint type defects within each consecutive time period for different historical defect type voiceprint data; obtaining the variation function of different state parameters within the corresponding time period for different voiceprint type defect time-domain variation functions, and determining the corresponding voiceprint type defect time-domain deconstancy variation function by combining the voiceprint-state normal relationship; and performing correlation analysis on defect parameters based on different voiceprint type defect time-domain deconstancy variation functions to form voiceprint type defect parameter correlation feature data.
[0016] In this invention, the feature analysis of acoustic fingerprint data for different defect types includes two aspects. First, it extracts corresponding characteristic acoustic fingerprint information from the missing acoustic fingerprint data of different types. Considering that defect acoustic fingerprint information itself contains some normal acoustic fingerprint information, normal acoustic fingerprint information can be removed based on the time-domain variation function of the defect acoustic fingerprint information. This removal requires obtaining the variation function of different state parameters of the blade within that time period, thus forming a clear time-domain variation function of the normal acoustic fingerprint for that period. Second, it is necessary to link the acoustic fingerprint information with the actual state of the defect. After removing the normal acoustic fingerprint information, the remaining acoustic fingerprint information is essentially only data related to the defect. This part of the acoustic fingerprint information is mainly affected by the state of the defect in the current time period. Since the state of the defect differs at different times, even exhibiting varying degrees of severity, it is necessary to deeply consider the relationship between the state of the defect and the acoustic fingerprint information, thereby providing important reference information for subsequently obtaining the specific situation of the defect through the acoustic fingerprint information.
[0017] As one possible implementation, correlation analysis of defect parameters is performed based on the time-domain deconstancy change function of different voiceprint type defects to form voiceprint type defect parameter correlation feature data. This includes: obtaining the defect characterization change function of defect characterization parameters that change with time within the corresponding time period for different voiceprint type defect time-domain deconstancy change functions; determining the corresponding time-period voiceprint type defect characterization correlation relationship by combining the corresponding defect characterization change function of different defect characterization parameters with the time-domain deconstancy change function of different voiceprint type defects; and determining the voiceprint type defect correlation relationship based on the voiceprint type defect characterization correlation relationship for different time periods.
[0018] In this invention, correlation analysis primarily utilizes the corresponding defect acoustic signature information to express the most representative time-varying defect parameter values. It should be noted that for a given defect type, there are multiple parameters representing different defect states, some of which change over time, while others remain fixed. For example, for a crack on a blade, if the crack length is chosen, it may remain essentially unchanged over a short period. However, if the resulting rotational axis runout or stress analysis is chosen, it changes over time. Either of these parameters can be chosen to directly characterize the specific crack condition. Therefore, to simplify the expression of the defect condition and to enable mapping with acoustic signature data, the selected defect characterization parameters must be time-varying and representative of the defect state. Of course, some parameters may not directly characterize the defect state, but are chosen as a compromise to meet the requirements of time-varying changes. This means that some parameters may be the same as those selected for other defect types. For example, selecting axial runout of the rotating shaft for a crack on a blade would be the same as selecting axial runout for a misaligned rotating shaft. However, this difference is because the essential reason for distinguishing between the two types of defects can be determined at the beginning based on the acoustic signature characteristics. The selected defect state parameters merely characterize the degree to which the state of this type of defect has reached, and will not cause misjudgment of the defect type. It should be noted that whether establishing correlation formulas or time-period acoustic signature-normal state formulas, they are all determined by directly performing a function operation on the parameters corresponding to one side of the equation that change over time. As for obtaining the final feature expression function using formulas obtained from different time periods, this is achieved by averaging the formulas determined at different time periods.
[0019] One possible approach is to acquire target voiceprint data and combine it with voiceprint defect identification feature data to perform defect identification analysis, thereby forming target voiceprint defect identification data. This includes: extracting the target voiceprint time-domain variation function and the target variation function of state parameters for different state parameters from the target voiceprint data; performing defect type identification analysis based on the target voiceprint time-domain variation function and voiceprint defect identification feature data to form target defect identification result information; and performing defect status confirmation analysis based on the target defect identification result information to form defect status analysis result data.
[0020] In this invention, after acquiring the voiceprint defect identification feature data, the feature data can be used to identify defects in the target voiceprint data. The identification includes two aspects: first, determining whether the target voiceprint data contains voiceprint information of a specific defect type and thus determining the defect type; second, after determining the defect type, it is necessary to confirm the current status of the defect in order to provide reference information for subsequent maintenance and processing.
[0021] As one possible implementation, based on the target voiceprint time-domain variation function and combined with voiceprint defect identification feature data, defect type identification analysis is performed to form target defect identification result information, including: determining the target voiceprint-state-normal relationship function based on different state parameter target variation functions and combined with the voiceprint-state-normal relationship formula; comparing the target voiceprint time-domain variation function with the target voiceprint-state-normal relationship function in the following way: if, within the target analysis period, the cumulative amount of the difference between the target voiceprint time-domain variation function and the target voiceprint-state-normal relationship function within the entire target analysis period does not exceed the voiceprint normal allowable... If the deviation is acceptable, a defect-free identification result is formed; if, within the target analysis period, the cumulative amount of the difference between the target voiceprint time-domain variation function and the target voiceprint-state normal relationship function exceeds the voiceprint normal allowable deviation throughout the entire target analysis period, then: the difference between the target voiceprint time-domain variation function and the target voiceprint-state normal relationship function is compared with the time-domain denormalization variation function of different voiceprint types, and the defect type corresponding to the minimum value of the difference between the target voiceprint time-domain variation function and the target voiceprint-state normal relationship function and the cumulative amount of the voiceprint type defect time-domain denormalization variation function within the target analysis period is determined as the target defect type.
[0022] In this invention, the identification of defect types first requires comparing the target acoustic print data with normal acoustic print data. Considering objective factors such as data acquisition and operating conditions, even if the target acoustic print data is normal, it is impossible for it to completely overlap with the characteristic function of normal acoustic print. Therefore, it is confirmed by comparing the normal allowable deviation of the acoustic print. If the deviation is within the allowable range, it can be determined that the target acoustic print information does not have defect features. If it exceeds the allowable range, it is confirmed that the blade has a defect. For the defect type that appears, it is determined by comparing the characteristic functions of different types of defects. Considering that the corresponding defect type will also have deviation, it is reasonable to select the defect type with the smallest cumulative deviation as the defect type reflected by the target acoustic print information.
[0023] As one possible implementation, based on the target defect identification results, a defect status confirmation analysis is performed to form defect status analysis result data, including: obtaining the corresponding voiceprint type defect correlation formula according to the determined defect type; and determining the target characterization parameter change function of the defect characterization parameters by combining the target voiceprint time domain change function with the target voiceprint-normal status relationship function.
[0024] In this invention, after determining the defect type, it is necessary to confirm the current defect status. Using the correlation formula corresponding to the defect type, the difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function is imported into the correlation formula as the defect-derived function. Following the model of the relationship formula, the variation function representing the parameters can be derived. With the help of this variation function, the important state parameters of the defect can be deduced, providing a specific reference for subsequent maintenance processing.
[0025] Secondly, the present invention provides a wind turbine blade acoustic signature defect identification system, comprising: a data acquisition unit for acquiring historical acoustic signature identification data and target acoustic signature data; a classification and division unit for classifying the historical acoustic signature identification data acquired by the data acquisition unit into defect types to form historical acoustic signature identification classification data; a feature extraction unit for extracting features for defect types from the historical acoustic signature identification classification data formed by the classification and division unit to form acoustic signature defect identification feature data; and a defect identification unit for identifying defects in the target acoustic signature data acquired by the data acquisition unit by combining the acoustic signature defect identification feature data formed by the feature extraction unit to form target acoustic signature defect identification data.
[0026] In this invention, the system combines different functional units to form a tightly integrated system capable of identifying defects in voiceprint data. The data acquisition unit continuously acquires historical data to expand the basic big data base; the classification and segmentation unit cleans the basic big data; the feature extraction unit extracts targeted features from the segmented big data; and the defect identification unit quickly and accurately identifies defects in the target voiceprint data, greatly improving analysis efficiency while ensuring the reliability and accuracy of defect identification.
[0027] The beneficial effects of the wind turbine blade acoustic defect identification method and system provided by this invention are as follows: This method acquires historical voiceprint recognition data to segment voiceprint data for different defect types and extract corresponding defect features, establishing a comprehensive defect identification database using big data. This database is then used to monitor the presence of defects in the data requiring defect identification in real time. On one hand, big data allows for accurate acquisition of characteristics of different types of defects, providing reliable and accurate comparative data for defect identification, thus ensuring accuracy. On the other hand, compared to traditional methods such as modal analysis, this method is more adaptable and can quickly and efficiently identify defects, significantly improving the real-time monitoring effect.
[0028] This system combines different functional units to form a tightly integrated system capable of identifying defects in voiceprint data. The data acquisition unit continuously acquires historical data to expand the basic big data; the classification and segmentation unit cleans the basic big data; the feature extraction unit extracts targeted features from the segmented big data; and the defect identification unit quickly and accurately identifies defects in the target voiceprint data, greatly improving analysis efficiency while ensuring the reliability and accuracy of defect identification. Attached Figure Description
[0029] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments of the present invention will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0030] Figure 1 This is a step diagram of the wind turbine blade acoustic defect identification method provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the structure of the wind turbine blade acoustic defect identification system provided in an embodiment of the present invention. Detailed Implementation
[0031] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention.
[0032] Wind turbine blades are the most important structural components of a wind turbine. Because they are constantly in motion and subjected to wind force, they are prone to failure, and maintenance after a failure is extremely inconvenient. Therefore, proactive monitoring and maintenance are necessary to improve wind turbine efficiency and reduce operating costs. Of course, wind turbine blades will exhibit operational defects before failure; real-time monitoring of these defects allows for effective proactive maintenance.
[0033] Currently, defect monitoring of wind turbine blades mainly includes manual inspection and intelligent data acquisition and analysis. Intelligent data acquisition and analysis primarily involves acquiring acoustic fingerprint data of wind turbine blades and performing a series of processes, including noise reduction and modal analysis, to determine whether defects exist based on the processed data. Although this method saves labor costs compared to manual inspection, it lacks real-time performance due to the complexity of the analysis methods, and the need for manual judgment of the analysis results makes the accuracy of defect identification unreliable.
[0034] refer to Figures 1-2 This invention provides a method for identifying acoustic fingerprint defects in wind turbine blades. This method acquires historical acoustic fingerprint data to segment the data for different defect types and extract corresponding defect features, establishing a comprehensive defect identification database using big data. This database is then used to monitor the presence of defects in the data requiring defect identification in real time. On one hand, big data allows for accurate acquisition of characteristics of different types of defects, providing reliable and accurate basic comparison data for defect identification, thus ensuring accuracy. On the other hand, compared to traditional methods such as modal analysis, this method is more adaptable and can quickly and efficiently identify defects, significantly improving the real-time monitoring effect of defect identification.
[0035] The method for identifying acoustic defects in wind turbine blades specifically includes the following steps: S1: Collect historical voiceprint recognition data, classify and categorize defect types to form historical voiceprint recognition classification data.
[0036] Historical voiceprint recognition data is collected and classified based on defect type to form historical voiceprint recognition classification data, including: extracting voiceprint data with normal monitoring results from historical voiceprint recognition data to form historical normal voiceprint data; extracting voiceprint data of different defect types from historical voiceprint recognition data to form corresponding historical defect type voiceprint data; and combining historical normal voiceprint data and historical defect type voiceprint data of different defect types to form historical voiceprint recognition classification data.
[0037] The classification of historical data mainly involves determining whether the voiceprint data has defects. The classification considers two aspects: firstly, extracting normal information data from historical data to extract features of normal voiceprint data, providing a positive reference for subsequent defect identification; secondly, extracting voiceprint data of different defect types to provide more targeted basic big data for subsequent feature extraction, ensuring the accuracy and rationality of voiceprint feature data for different defect types.
[0038] S2: Based on historical voiceprint recognition classification data, feature extraction is performed for defect types to form voiceprint defect recognition feature data.
[0039] Based on historical voiceprint recognition and classification data, feature extraction is performed for defect types to form voiceprint defect recognition feature data. This includes: performing feature analysis on blade state based on historical normal voiceprint data to form normal voiceprint feature data; combining normal voiceprint feature data to perform defect feature analysis on voiceprint data of different historical defect types to form defect voiceprint feature data corresponding to different defect types; and combining normal voiceprint feature data and different defect voiceprint feature data to form voiceprint defect recognition feature data.
[0040] Voiceprint feature extraction includes two aspects: one is the extraction of normal voiceprint features, which can be obtained directly from historical normal voiceprint big data; the other is the extraction of defect voiceprint feature data for different defect types. Since the voiceprint features exhibited by different defect types are different, and defects are only a small part of the entire voiceprint information, most of the voiceprint data is still normal information. Therefore, when conducting analysis, it is necessary to consider combining normal voiceprint feature data to assist in the extraction of defect voiceprint features.
[0041] Based on historical normal acoustic signature data, feature analysis is performed on the blade condition to form normal acoustic signature feature data, including: obtaining the normal temporal domain variation function of acoustic signature for each consecutive time period based on historical normal acoustic signature data. Where n represents the number of different consecutive time periods; for different voiceprint normal time-domain variation functions Obtain the normal state parameter change function for different blade state parameters. Where m represents the number of different state parameters; for different voiceprint normal time-domain variation functions According to the corresponding normal state parameter change function Establish the corresponding time period voiceprint-normal status relationship. According to the normal relationship between voiceprint and status in different time periods The relationship between voiceprint and normal status was determined. ,in, .
[0042] For normal acoustic fingerprint data, the main factors affecting its characteristics are the blade's operating state, such as rotational speed, acceleration, wind resistance, temperature, and humidity. Therefore, to obtain the characteristic information of normal acoustic fingerprints, the influence of these state parameters on the acoustic fingerprint features should be fully considered. First, normal acoustic fingerprint data for each consecutive time period is acquired to determine the correlation between different state parameters and the temporal variation function of the acoustic fingerprint in the time dimension under multiple sample data items. Considering that the acquired normal acoustic fingerprint data, although denoised, may still be affected by noise and may be influenced by other factors, leading to differences in the correlation formula obtained at different time periods, an averaging function is used to determine the correlation characteristic formula of normal acoustic fingerprints. Although the established correlation formula lacks accuracy in expressing the relationship between acoustic fingerprint information and state parameters, it can serve as important reference data for eliminating interference from normal acoustic fingerprint information during subsequent defect feature extraction.
[0043] By combining normal voiceprint feature data, defect feature analysis is performed on voiceprint data of different historical defect types to form defect voiceprint feature data corresponding to different defect types. This includes: obtaining the time-domain variation function of voiceprint type defects within each consecutive time period for different historical defect type voiceprint data. Where k represents the number of different defect types, and i represents the number of different consecutive time periods under defect type k; the time-domain variation function for different voiceprint defect types. Obtain the change function of different state parameters within the corresponding time period, and combine it with the voiceprint-state normal relationship. The time-domain deconstancy function of the corresponding voiceprint type defect was determined. ,in, Based on the time-domain denormalization function of different voiceprint types... Correlation analysis was performed on the defect parameters to generate correlation feature data of voiceprint type defect parameters.
[0044] Feature analysis of acoustic fingerprint data for different defect types includes two aspects. First, it involves extracting corresponding characteristic acoustic fingerprint information from the missing data of different types. Considering that defect acoustic fingerprint information itself contains some normal acoustic fingerprint information, normal acoustic fingerprint information can be removed based on the time-domain variation function of the defect acoustic fingerprint information. This removal requires obtaining the variation functions of different state parameters of the blade within that time period, thus forming a clear time-domain variation function of the normal acoustic fingerprint for that period. Second, it is necessary to correlate the acoustic fingerprint information with the actual state of the defect. After removing the normal acoustic fingerprint information, the remaining acoustic fingerprint information is essentially only data related to the defect. This part of the acoustic fingerprint information is mainly affected by the state of the defect at the current time period. Since the state of the defect varies at different times, even differing in severity, it is necessary to deeply consider the relationship between the state of the defect and the acoustic fingerprint information. This provides important reference information for subsequently obtaining the specific situation of the defect through the acoustic fingerprint information.
[0045] Based on the time-domain denormalization function of different voiceprint types Correlation analysis was performed on the defect parameters to generate correlation feature data for voiceprint type defect parameters, including: time-domain constant variation functions for different voiceprint type defects. Obtain the defect characterization change function of the defect characterization parameters as a function of time within the corresponding time period. ; Temporal deconsolidation function for different voiceprint type defects Defect characterization change function combined with different corresponding defect characterization parameters The correlation formula for the characteristic features of voiceprint type defects in the corresponding time period was determined. Correlation relationship of voiceprint type defects in different time periods The correlation formula of voiceprint type defects was determined. ,in, .
[0046] Correlation analysis primarily utilizes the acoustic signature information of corresponding defects to express the most representative time-varying defect parameter values. It's important to note that for a given defect type, there are multiple parameters representing different defect states, some of which change over time, while others remain constant. For example, for a crack on a blade, choosing the crack length might result in a relatively short period of time that remains essentially unchanged, while choosing the resulting rotational axis runout or stress analysis would show that these parameters change over time. Either of these parameters can be chosen to directly characterize the specific crack condition. Therefore, to simplify the expression of defect conditions and to facilitate mapping with acoustic signature data, the selected defect characterization parameters must be time-varying and representative of the defect state. Of course, some parameters may not directly characterize the defect state, but are chosen as a compromise to meet the requirements of time-varying changes. This means that some parameters may be the same as those selected for other defect types. For example, selecting axial runout of the rotating shaft for a crack on a blade would be the same as selecting axial runout for a misaligned rotating shaft. However, this difference is because the essential reason for distinguishing between the two types of defects can be determined at the beginning based on the acoustic signature characteristics. The selected defect state parameters merely characterize the degree to which the state of this type of defect has reached, and will not cause misjudgment of the defect type. It should be noted that whether establishing correlation formulas or time-period acoustic signature-normal state formulas, they are all determined by directly performing a function operation on the parameters corresponding to one side of the equation that change over time. As for obtaining the final feature expression function using formulas obtained from different time periods, this is achieved by averaging the formulas determined at different time periods.
[0047] S3: Acquire target voiceprint data and combine it with voiceprint defect recognition feature data to perform defect recognition analysis and form target voiceprint defect recognition data.
[0048] Acquire target voiceprint data and combine it with voiceprint defect recognition feature data to perform defect identification analysis, forming target voiceprint defect identification data. This includes: extracting the target voiceprint time-domain variation function and the target variation function of state parameters for different states based on the target voiceprint data; performing defect type identification analysis based on the target voiceprint time-domain variation function and voiceprint defect recognition feature data to form target defect identification result information; and performing defect status confirmation analysis based on the target defect identification result information to form defect status analysis result data.
[0049] After obtaining the voiceprint defect identification feature data, the feature data can be used to identify defects in the target voiceprint data. The identification includes two aspects: first, determining whether the target voiceprint data contains voiceprint information of a specific defect type and thus determining the defect type; second, after determining the defect type, it is necessary to confirm the current status of the defect in order to provide reference information for later maintenance and processing.
[0050] Based on the target voiceprint time-domain variation function, and combined with voiceprint defect identification feature data, defect type identification analysis is performed to form target defect identification result information, including: target variation function based on different state parameters, combined with voiceprint-state normal relationship formula. Determine the target voiceprint-normal state relationship function. ; the time-domain variation function of the target voiceprint Same target voiceprint - normal state relationship function Compare the following methods: if within the target analysis cycle Within this range, the cumulative difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function over the entire target analysis period does not exceed the allowable deviation of the voiceprint normality. This results in a defect-free identification result, where, This indicates the normal allowable deviation for voiceprint analysis; if within the target analysis cycle... Within this range, the cumulative difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function exceeds the allowable deviation of the voiceprint normality throughout the entire target analysis period, i.e. Then: the difference between the target acoustic signature time-domain variation function and the target acoustic signature-normal state relationship function is compared with the time-domain deconstancy variation functions of different acoustic signature types. The defect type corresponding to the minimum value of the difference between the target acoustic signature time-domain variation function and the target acoustic signature-normal state relationship function and the cumulative value of the time-domain deconstancy variation function of the acoustic signature type defect within the target analysis period is determined as the target defect type. Different from different Compare and select the smallest. The corresponding defect type is determined as the target defect type.
[0051] The identification of defect types first requires comparing the target acoustic print data with normal acoustic print data. Considering objective factors such as data acquisition and operating conditions, even if the target acoustic print data is normal, it is impossible for it to completely overlap with the characteristic function of normal acoustic prints. Therefore, it is confirmed by comparing the normal allowable deviation of the acoustic print. If the deviation is within the allowable range, it can be determined that the target acoustic print information does not show defect features. If it exceeds the allowable range, it is confirmed that there is a defect in the blade. For the defect type that appears, it is determined by comparing the characteristic functions of different types of defects. Considering that the corresponding defect type will also have deviation, it is reasonable to select the defect type with the smallest cumulative deviation as the defect type reflected by the target acoustic print information.
[0052] Based on the target defect identification results, a defect status confirmation analysis is performed to generate defect status analysis result data, including: obtaining the corresponding voiceprint type defect correlation formula based on the determined defect type. Based on the target voiceprint time-domain variation function Same target voiceprint - normal state relationship function and combined The target characterization parameter variation function of the defect characterization parameters is determined.
[0053] After determining the defect type, it is necessary to confirm the current defect status. Using the correlation formula corresponding to the defect type, the difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function is imported into the correlation formula as the defect-derived function. Following the model of the relationship formula, the variation function representing the parameters can be derived. With the help of this variation function, the important state parameters of the defect can be deduced, providing a specific reference for subsequent maintenance.
[0054] The present invention also provides a wind turbine blade acoustic signature defect identification system, which includes: a data acquisition unit for acquiring historical acoustic signature identification data and target acoustic signature data; a classification and division unit for classifying the historical acoustic signature identification data acquired by the data acquisition unit into defect types to form historical acoustic signature identification classification data; a feature extraction unit for extracting features for defect types from the historical acoustic signature identification classification data formed by the classification and division unit to form acoustic signature defect identification feature data; and a defect identification unit for identifying defects in the target acoustic signature data acquired by the data acquisition unit by combining the acoustic signature defect identification feature data formed by the feature extraction unit to form target acoustic signature defect identification data.
[0055] This system combines different functional units to form a tightly integrated system capable of identifying defects in voiceprint data. The data acquisition unit continuously acquires historical data to expand the basic big data; the classification and segmentation unit cleans the basic big data; the feature extraction unit extracts targeted features from the segmented big data; and the defect identification unit quickly and accurately identifies defects in the target voiceprint data, greatly improving analysis efficiency while ensuring the reliability and accuracy of defect identification.
[0056] In summary, the beneficial effects of the wind turbine blade acoustic defect identification method and system provided in this embodiment of the invention are as follows: This method acquires historical voiceprint recognition data to segment voiceprint data for different defect types and extract corresponding defect features, establishing a comprehensive defect identification database using big data. This database is then used to monitor the presence of defects in the data requiring defect identification in real time. On one hand, big data allows for accurate acquisition of characteristics of different types of defects, providing reliable and accurate comparative data for defect identification, thus ensuring accuracy. On the other hand, compared to traditional methods such as modal analysis, this method is more adaptable and can quickly and efficiently identify defects, significantly improving the real-time monitoring effect.
[0057] This system combines different functional units to form a tightly integrated system capable of identifying defects in voiceprint data. The data acquisition unit continuously acquires historical data to expand the basic big data; the classification and segmentation unit cleans the basic big data; the feature extraction unit extracts targeted features from the segmented big data; and the defect identification unit quickly and accurately identifies defects in the target voiceprint data, greatly improving analysis efficiency while ensuring the reliability and accuracy of defect identification.
[0058] In the embodiments of this application, "instruction" can include direct and indirect instructions, as well as explicit and implicit instructions. The information indicated by a certain piece of information is called the information to be instructed. In the specific implementation process, there are many ways to instruct the information to be instructed, such as, but not limited to, directly instructing the information to be instructed, such as the information to be instructed itself or its index. It can also indirectly instruct the information to be instructed by instructing other information, where there is a relationship between the other information and the information to be instructed. It can also instruct only a part of the information to be instructed, while the other parts are known or pre-agreed upon. For example, the instruction of specific information can be achieved by using a pre-agreed (e.g., protocol-defined) arrangement of various pieces of information, thereby reducing instruction overhead to some extent. At the same time, common parts of various pieces of information can be identified and uniformly indicated to reduce the instruction overhead caused by individually indicating the same information.
[0059] Furthermore, the specific indication method can also be any existing indication method, such as, but not limited to, the above-mentioned indication methods and their various combinations. Specific details of various indication methods can be found in existing technologies, and will not be repeated here. As described above, for example, when multiple pieces of information of the same type need to be indicated, the indication methods for different pieces of information may differ. In the specific implementation process, the required indication method can be selected according to specific needs. This application embodiment does not limit the selected indication method; therefore, the indication methods involved in this application embodiment should be understood to cover various methods that enable the party to be indicated to obtain the information to be indicated.
[0060] It should be understood that the information to be indicated can be sent as a whole or divided into multiple sub-information messages sent separately, and the sending period and / or timing of these sub-information messages can be the same or different. The specific sending method is not limited in this application embodiment. The sending period and / or timing of these sub-information messages can be predefined, for example, according to a protocol, or configured by the sending device by sending configuration information to the receiving device.
[0061] "Predefined" or "pre-configured" can be achieved by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in the device. This application does not limit the specific implementation method. "Saving" can refer to saving in one or more memories. These memories can be separate installations or integrated into the encoder, decoder, processor, or communication device. Alternatively, some memories can be separately installed, while others are integrated into the decoder, processor, or communication device. The type of memory can be any form of storage medium, and this application does not limit this.
[0062] The “protocol” mentioned in the embodiments of this application may refer to a protocol family in the field of communication, a standard protocol with a similar protocol family frame structure, or a related protocol applied to future communication systems. The embodiments of this application do not specifically limit this.
[0063] In the embodiments of this application, descriptions such as "when," "under the circumstances," "if," and "if" all refer to the device making corresponding processing under certain objective circumstances, and are not limited to a specific time. They do not require the device to make a judgment action during implementation, nor do they imply any other limitations.
[0064] In the description of the embodiments of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can represent A or B. "And / or" in the embodiments of this application is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, and B alone, where A and B can be singular or plural. Furthermore, in the description of the embodiments of this application, unless otherwise stated, "multiple" refers to two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple. Additionally, to facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with essentially the same function and effect. Those skilled in the art will understand that the terms "first," "second," etc., do not limit the quantity or order of execution, and that "first," "second," etc., are not necessarily different. Furthermore, in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate that something is being used as an example, illustration, or description. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner for ease of understanding.
[0065] It should be understood that the processor in the embodiments of this application can be a central processing unit (CPU), or it can be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor, etc.
[0066] It should also be understood that the memory in the embodiments of this application can be volatile memory or non-volatile memory, or may include both volatile and non-volatile memory. The non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. The volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of random access memory (RAM) are available, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate synchronous DRAM (DDR SDRAM), enhanced synchronous DRAM (ESDRAM), synchronous linked DRAM (SLDRAM), and direct rambus RAM (DR RAM).
[0067] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.
[0068] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.
[0069] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be a single item or multiple items.
[0070] It should be understood that in the various embodiments of this application, the order of the above-mentioned processes does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.
[0071] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0072] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0073] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.
[0074] The units described as separate components may or may not be physically separate. The components shown as units 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 units can be selected to achieve the purpose of this embodiment according to actual needs.
[0075] In addition, the functional units 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.
[0076] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0077] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
Claims
1. A method for identifying acoustic defects in wind turbine blades, characterized in that, include: Collect historical voiceprint recognition data, classify the defect types, and form historical voiceprint recognition classification data; Based on the historical voiceprint recognition classification data, feature extraction is performed for the defect type to form voiceprint defect recognition feature data. The target voiceprint data is acquired and combined with the voiceprint defect identification feature data to perform defect identification analysis, thereby forming target voiceprint defect identification data.
2. The method for identifying acoustic defects in wind turbine blades according to claim 1, characterized in that, The collected historical voiceprint recognition data is classified based on defect type to form historical voiceprint recognition classification data, including: Based on the historical voiceprint recognition data, voiceprint data with normal monitoring results are extracted to form historical normal voiceprint data; Based on the historical voiceprint recognition data, voiceprint data of different defect types are extracted to form corresponding historical defect type voiceprint data. The historical normal voiceprint data and the historical defect type voiceprint data of different defect types are combined to form the historical voiceprint recognition and classification data.
3. The method for identifying acoustic defects in wind turbine blades according to claim 2, characterized in that, The step of extracting features based on the historical voiceprint recognition classification data to form voiceprint defect recognition feature data includes: Based on the historical normal voiceprint data, feature analysis is performed on the blade condition to form normal voiceprint feature data; By combining the normal voiceprint feature data, defect feature analysis is performed on the voiceprint data of different historical defect types to form defect voiceprint feature data corresponding to different defect types. The normal voiceprint feature data and the different defective voiceprint feature data are combined to form the voiceprint defect identification feature data.
4. The method for identifying acoustic defects in wind turbine blades according to claim 2, characterized in that, The step of performing feature analysis on the blade condition based on the historical normal voiceprint data to form normal voiceprint feature data includes: Based on the historical normal voiceprint data, obtain the normal time-domain variation function of the voiceprint in each consecutive time period; For different acoustic signature normal time-domain variation functions, obtain the corresponding normal state parameter variation functions for different blade state parameters; For different normal time-domain variation functions of the voiceprint, establish corresponding time-period voiceprint-normal state relationship based on the corresponding different normal state parameter variation functions; Based on the different time periods, the voiceprint-state normal relationship is determined.
5. The method for identifying acoustic defects in wind turbine blades according to claim 4, characterized in that, The step involves combining the normal voiceprint feature data with the defect feature analysis of different historical defect type voiceprint data to form defect voiceprint feature data corresponding to different defect types, including: For different types of historical defect voiceprint data, obtain the time-domain variation function of the voiceprint type defect in each consecutive time period; For different types of voiceprint defects, the change functions of different state parameters within the corresponding time period are obtained, and combined with the voiceprint-state normal relationship, the corresponding voiceprint type defect time-domain constant change function is determined. Correlation analysis of defect parameters is performed based on the time-domain constant change function of different voiceprint type defects to form voiceprint type defect parameter correlation feature data.
6. The method for identifying acoustic defects in wind turbine blades according to claim 5, characterized in that, The correlation analysis of defect parameters based on the time-domain deconstancy function of different voiceprint types is performed to form voiceprint type defect parameter correlation feature data, including: For different types of soundprint defects, the time-domain deconstancy function is used to obtain the defect characterization parameter change function that varies with time within the corresponding time period. For different types of voiceprint defects, the time-domain constant change function is combined with the defect characterization change function of the corresponding different defect characterization parameters to determine the corresponding time-period voiceprint type defect characterization correlation formula. Based on the correlation relationship of voiceprint type defect characterization for different time periods, the correlation relationship of voiceprint type defect is determined.
7. The method for identifying acoustic defects in wind turbine blades according to claim 6, characterized in that, The process of acquiring target voiceprint data and combining it with voiceprint defect identification feature data to perform defect identification analysis and form target voiceprint defect identification data includes: Based on the target voiceprint data, extract the target voiceprint time-domain variation function and the target variation function of state parameters with different state parameters; Based on the target voiceprint time-domain variation function, and combined with the voiceprint defect identification feature data, defect type identification analysis is performed to form target defect identification result information; Based on the target defect identification results, a defect status confirmation analysis is performed to generate defect status analysis result data.
8. The method for identifying acoustic defects in wind turbine blades according to claim 7, characterized in that, The step of identifying and analyzing the defect type based on the target voiceprint time-domain variation function and the voiceprint defect identification feature data to form target defect identification result information includes: Based on the different target change functions of the state parameters, and in combination with the voiceprint-state normal relationship formula, the target voiceprint-state normal relationship function is determined; The target voiceprint time-domain variation function is compared with the target voiceprint-normal state relationship function in the following manner: If, within the target analysis period, the cumulative difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function does not exceed the normal allowable deviation of the voiceprint during the entire target analysis period, then a defect-free identification result is formed. If, within the target analysis period, the cumulative difference between the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function exceeds the allowable deviation of voiceprint normality throughout the entire target analysis period, then: The difference between the target voiceprint time-domain variation function and the target voiceprint-state normal relationship function is compared with the time-domain deconstancy variation functions of different voiceprint types. The defect type corresponding to the minimum value of the difference between the target voiceprint time-domain variation function and the target voiceprint-state normal relationship function and the time-domain deconstancy variation function of the voiceprint type defect in the target analysis period is determined as the target defect type.
9. The method for identifying acoustic defects in wind turbine blades according to claim 8, characterized in that, The step of performing a defect status confirmation analysis based on the target defect identification result information to form defect status analysis result data includes: Based on the determined defect type, obtain the corresponding correlation formula for the voiceprint type defect; Based on the target voiceprint time-domain variation function and the target voiceprint-normal state relationship function, and combined with the voiceprint type defect correlation formula, the target characterization parameter variation function of the defect characterization parameter is determined.
10. A wind turbine blade acoustic signature defect identification system, employing the wind turbine blade acoustic signature defect identification method according to any one of claims 1-9, characterized in that, include: The data acquisition unit is used to acquire historical voiceprint recognition data and target voiceprint data; The classification and division unit is used to classify the historical voiceprint recognition data acquired by the data acquisition unit according to the defect type, and form historical voiceprint recognition classification data. The feature extraction unit is used to extract features for defect types from the historical voiceprint recognition classification data formed by the classification and division unit to form voiceprint defect recognition feature data. The defect identification unit is used to identify defects in the target voiceprint data acquired by the data acquisition unit by combining the voiceprint defect identification feature data formed by the feature extraction unit, thereby forming target voiceprint defect identification data.