Blade connection state evaluation method and device and readable storage medium

Abstract

The invention discloses a blade connection state evaluation method and device and a readable storage medium, and the method comprises the steps: periodically and synchronously collecting multiple acoustic signals of multiple blades, so as to extract multiple denoising signals corresponding to the multiple acoustic signals, and each acoustic signal corresponds to one blade; impact components in the de-noised signals are extracted, impact energy factors corresponding to the de-noised signals are calculated, and the impact energy factors represent the energy intensity of the impact components; and for any path of target de-noised signal in the paths of de-noised signals, calculating an abnormal value proportion of an impact energy factor corresponding to the target de-noised signal, and evaluating a connection state of a blade corresponding to the target de-noised signal based on the abnormal value proportion. According to the technical scheme provided by the invention, the blade connection state can be monitored at relatively low cost.

Description

Technical Field

[0001] This application relates to the field of acoustic monitoring technology, and in particular to a method, apparatus and readable storage medium for evaluating the blade connection status. Background Technology

[0002] Blade root bolts are critical load-bearing components connecting wind turbine blades to the hub, and their structural integrity directly affects the operational safety of the wind turbine. Loose or broken bolts can easily lead to blade detachment or even complete flight. This not only causes severe damage to the wind turbine but also poses a serious threat to personnel safety. Therefore, monitoring the connection status of blade root bolts is of paramount importance.

[0003] However, existing methods for monitoring the condition of bolted connections are typically direct monitoring methods, which involve placing strain gauges, ultrasonic sensors, or displacement sensors on each bolt to monitor its axial stress, preload, or fretting displacement in real time. While this method can monitor the bolted connection condition with high precision, it is costly and involves complex installation procedures.

[0004] Therefore, it is necessary to provide a new method and apparatus for evaluating the blade connection status to address the above-mentioned shortcomings. Summary of the Invention

[0005] The purpose of this application is to provide a method, apparatus and readable storage medium for evaluating the blade connection status, which can monitor the blade connection status at a lower cost.

[0006] To achieve the above objectives, this application provides a method for evaluating the connection status of a blade. The method is applied to a wind turbine generator having multiple blades. The method includes: periodically and synchronously acquiring multiple acoustic signals from the multiple blades to extract multiple denoised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade; extracting the impact component from each denoised signal to calculate the impact energy factor corresponding to each denoised signal, wherein the impact energy factor characterizes the energy intensity of the impact component; and for any target denoised signal among the denoised signals, calculating the outlier percentage of the impact energy factor corresponding to the target denoised signal, and evaluating the connection status of the blade corresponding to the target denoised signal based on the outlier percentage.

[0007] To achieve the above objectives, this application also provides a blade connection status assessment device, which is applied to a wind turbine having multiple blades. The device includes: a signal acquisition module for periodically and synchronously acquiring multiple acoustic signals from the multiple blades to extract multiple denoised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade; a signal processing module for extracting the impact component from each of the denoised signals to calculate the impact energy factor corresponding to each of the denoised signals, wherein the impact energy factor characterizes the energy intensity of the impact component; and a blade assessment module for calculating the outlier percentage of the impact energy factor corresponding to any one of the target denoised signals, and assessing the connection status of the blade corresponding to the target denoised signal based on the outlier percentage.

[0008] To achieve the above objectives, this application also provides a blade connection status assessment device, comprising: a memory for storing a computer program; and a processor for executing the computer program stored in the memory, so that the device performs the blade connection status assessment method as described above.

[0009] To achieve the above objectives, this application also provides a computer-readable storage medium having instructions stored thereon, which, when executed by a processor, cause the processor to implement the blade connection state evaluation method as described above.

[0010] To achieve the above objectives, this application also provides a computer program product comprising computer program code, which, when run on a computer, enables the computer to implement the blade connection status evaluation method as described above.

[0011] To achieve the above objectives, this application also provides a chip that includes a circuit for performing the blade connection status evaluation method as described above.

[0012] Therefore, the technical solution provided in this application first periodically and synchronously acquires multiple acoustic signals from multiple blades of a wind turbine. Then, it utilizes the structural symmetry of the multiple blades to estimate and suppress common-mode acoustic background, thereby converting each acoustic signal into a corresponding denoised signal. Next, leveraging the sparsity of structural transient anomalies in the time-frequency domain, it extracts the impact component characterizing transient acoustic features from the denoised signal to calculate the corresponding impact energy factor. Based on a sliding time window, it fuses and models all impact energy factors to generate a dynamically updated anomaly detection threshold. Finally, by statistically analyzing the proportion of anomaly values ​​in any acoustic signal's impact energy factor, it achieves a quantitative assessment of the blade root connection status. This solution, through common-mode background estimation and suppression of multi-channel acoustic signals, can effectively distinguish between local structural anomalies and environmental interference in high-noise environments using only acoustic sensors. This significantly improves the reliability of bolt loosening or breakage fault identification and allows for monitoring of blade connection status at a lower cost. Attached Figure Description

[0013] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0014] Figure 1 This is a flowchart of a blade connection status evaluation method according to one embodiment of this application; Figure 2 This is a trend diagram of the impact energy factor of a fan in one embodiment of this application; Figure 3 This is a schematic diagram of the functional modules of the blade connection status evaluation device in one embodiment of this application; Figure 4 This is a schematic diagram of the blade connection status evaluation device according to one embodiment of this application. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be further described in detail below with reference to the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. In the description of this application, the terms "first," "second," "third," etc., are only used to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances. Furthermore, in the description of this application, unless otherwise stated, when an element is described as "connected" to another element, it can be directly connected to the other element, or there can be one or more intermediate elements between them. "Multiple" refers to two or more. "And / or" describes 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, and B existing alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.

[0016] During long-term operation, the connection between the blades and hub of a wind turbine is prone to loosening or even breakage due to fatigue, corrosion, or installation defects. The blade root bolts are critical load-bearing components connecting the blades and hub, and their structural integrity directly affects the operational safety of the wind turbine. Loosening or breakage of these bolts can easily lead to blade detachment or even fly-off accidents. This not only causes severe damage to the wind turbine but also poses a serious threat to personnel safety. As wind turbines develop towards larger megawatts and longer blades, the load conditions on the blade root bolts are becoming increasingly complex, significantly increasing their potential failure risk. Therefore, monitoring the connection status of the blades is of significant engineering importance.

[0017] Existing methods for monitoring the connection status of blades are mainly divided into two categories. The first category is the direct monitoring method, which monitors the axial stress, preload, or fretting displacement in real time by placing strain gauges, ultrasonic sensors, or displacement sensors on each bolt. Although this method can monitor the bolt connection status with high precision, the sensors are expensive, the installation process is complex, and maintenance is difficult. It is also difficult to deploy on a large scale in the densely bolted blade root area, resulting in poor engineering economy and feasibility. The second category is the non-contact monitoring method, which uses inductive sensors, lidar, cameras, etc., to count bolts or compare them with preset points to determine the presence of bolts, thereby achieving visual monitoring of bolt loosening or breakage. This type of method is susceptible to interference from environmental factors such as rain, snow, oil, and changes in light, resulting in high false alarm and false negative rates.

[0018] Therefore, how to provide a blade connection status monitoring method that is both reliable and economical has become an urgent issue to be addressed in this field.

[0019] The present application will now be described in more detail with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are for illustrative and explanatory purposes only and do not constitute a limitation on the embodiments of the present application. The embodiments described herein are only a part of the embodiments of the present application, and not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of this application.

[0020] The applicant discovered that the multiple blades of a wind turbine exhibit high symmetry in geometry, mass distribution, and installation layout. This means that when the wind turbine operates under uniform wind load, the aerodynamic load on each blade is approximately equal, resulting in highly synchronized structural vibrations (such as blade bending and torsional vibrations) in frequency, amplitude, and phase. Furthermore, during normal operation, the coupled vibration transmission paths between each blade and components such as the hub and gearbox are consistent. These characteristics cause acoustic signals such as wind noise, gearbox noise, and generator noise collected at similar locations on the blade structure to exhibit highly correlated common-mode characteristics. However, when a blade suffers localized damage such as loose bolts or breakage, the resulting transient acoustic emission signals will exhibit differential-mode characteristics in specific channels. Based on this physical mechanism, the applicant discovered that environmental interference and local fault signals can be separated by common-mode background estimation and suppression, and by leveraging the spatial inconsistency of abnormal responses among multiple channels, the identification and location of local faults in a single blade can be achieved.

[0021] The solution in this application is applicable to wind turbines having multiple blades, with each blade evenly arranged circumferentially around the hub. Please refer to [link / reference]. Figure 1 , Figure 1 This is a flowchart of a blade connection status evaluation method according to one embodiment of this application.

[0022] S101: Periodically and synchronously acquire multiple acoustic signals from the multiple blades to extract multiple denoised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade.

[0023] In this embodiment, the system can periodically and synchronously acquire acoustic signals from each blade of the wind turbine, thereby forming a set of time-aligned, multi-channel audio segments with fixed durations. Each audio segment (i.e., acoustic signal) corresponds to one blade. Specifically, operators can deploy one or more acoustic sensors at the same structural location on each blade (e.g., at the blade root baffle, blade main sparsity cap, etc.). These acoustic sensors are used to acquire the local acoustic signals of the corresponding blade in real time. After the deployment of the acoustic sensors is completed, the system can synchronously acquire the signals from each acoustic sensor with a fixed step size based on a unified time reference, thereby obtaining time-aligned signals from multiple channels (i.e., multiple acoustic signals).

[0024] In one implementation, the system can synchronously trigger the acquisition of acoustic sensors on each blade at fixed time intervals based on a unified clock reference, with each acquisition lasting a fixed duration, thereby directly generating a set of multi-channel audio segments with fixed durations and time alignment. In another implementation, the system can also continuously record the acoustic channel corresponding to each blade, forming multiple continuous audio streams. Then, based on a unified time reference (such as the system clock), the audio streams are synchronously sliced ​​at fixed time intervals and fixed window lengths to obtain a set of multi-channel audio segments with fixed durations and time alignment.

[0025] For ease of explanation, this application uses a wind turbine with three blades (denoted as blade A, blade B, and blade C) as an example. After synchronously acquiring continuous acoustic signals from the roots of the three blades, the system can divide each audio stream into a series of independent audio segments of uniform duration according to a fixed period, and store them in a standard audio file format containing timestamp information. For example, the system can extract three audio streams acquired within the same time window into audio segments of fixed duration (e.g., 60 seconds) and record them as... ,in, The acoustic signal is from the root of leaf A. This is the acoustic signal at the root of leaf B. This represents the acoustic signal at the root of leaf C. It should be noted that... Each consists of a set of audio segments arranged in chronological order.

[0026] Once the system acquires the aforementioned multiple acoustic signals, it can perform noise reduction processing on each of the multiple acoustic signals, removing interference signals such as wind noise, gearbox noise, and generator noise from the original acoustic signals, and converting each acoustic signal into a corresponding noise-reduced signal, thereby extracting a multi-channel denoised signal that is consistent with the original multiple acoustic signals in terms of the number of channels and the time dimension.

[0027] In one implementation, the extraction of the multi-channel denoised signal corresponding to the multi-channel acoustic signal can be achieved in the following way: First, the common-mode acoustic background signal of the multiple acoustic signals is calculated; then, the common-mode acoustic background signal is removed from each acoustic signal to obtain the multi-channel denoised signal corresponding to the multiple acoustic signals.

[0028] Specifically, due to the high symmetry in the geometry, mass distribution, and installation layout of the blades of a wind turbine, the aerodynamic load on each blade is approximately equal when the wind turbine is operating under uniform wind load. This results in highly synchronized structural vibrations (such as blade bending and torsional vibrations) in frequency, amplitude, and phase. Furthermore, during normal operation, the coupled vibration transmission paths between each blade and components such as the hub, gearbox, and generator are consistent. This type of noise is a radially propagating uniform field; therefore, it can be assumed that each acoustic sensor receives background noise of similar intensity. This background noise can be considered a common-mode acoustic background signal, and thus, initial elimination of background noise can be achieved through common-mode suppression.

[0029] When calculating the common-mode acoustic background signal of multiple acoustic signals, the system first acquires the sampled values ​​of each acoustic signal at the same time, and then performs common-mode estimation on each sampled value to obtain the common-mode acoustic background signal. In other words, at each sampling time, the system first acquires the synchronous sampled values ​​of all channel acoustic signals at that time, and then performs common-mode estimation on this set of sampled values ​​to generate the common-mode acoustic background signal point by point. It should be noted that the common-mode estimation operation includes at least one of arithmetic mean, median filtering, or weighted average.

[0030] Continuing with the example of blades A, B, and C, we will analyze the three sets of acoustic signals collected simultaneously. and By performing a weighted average, the common-mode acoustic background signal can be obtained. An approximate estimate. Therefore, the common-mode acoustic background signal generated by the mechanical vibration of components such as gearboxes and generators. The following formula can be used to estimate:

[0031] After obtaining the common-mode acoustic background signal Then, the system can use this as a noise reference, and then employ denoising algorithms such as time-domain or frequency-domain spectral subtraction, Wiener filtering, and adaptive filtering to denoise the aforementioned multi-channel acoustic signals. Remove common-mode acoustic background signal The interference was eliminated to obtain the corresponding multi-channel denoised signal. .

[0032] It should be noted that the denoised signal And the denoised signal Each of them consists of a set of audio segments arranged in chronological order.

[0033] S102: Extract the impulse component from each of the denoised signals to calculate the impulse energy factor corresponding to each of the denoised signals, wherein the impulse energy factor characterizes the energy intensity of the impulse component.

[0034] In this embodiment, after obtaining multiple denoised signals... Subsequently, based on the characteristic that when the blade suffers local damage such as bolt loosening or breakage, the transient acoustic emission signal generated will exhibit differential mode characteristics of a specific channel, the system can extract the differential mode characteristic signal (i.e., the impact component) contained in the above-mentioned denoised signal.

[0035] In one feasible implementation, the system first performs time-frequency transformation on each denoised signal to obtain a time spectrum corresponding to each denoised signal. Then, it utilizes the sparsity of the impulse component in the time-frequency domain and, based on the structural sparsity of each time spectrum, uses sparse decomposition or robust principal component analysis to separate the impulse component corresponding to the denoised signal from each time spectrum. The impulse component is used to characterize the transient acoustic features in the denoised signal (e.g., the sound signal emitted when a bolt loosens or breaks).

[0036] In another feasible implementation, the system can also employ the HPSS (Harmonic-Percussive Source Separation) algorithm to decompose the denoised signal into harmonic and impulse components. The denoised signal... For example, the system can utilize the short-time Fourier transform to... The complex spectrum is obtained by transforming to the time-frequency domain. Then, the harmonic components are separated based on the geometric differences between the harmonics and the impulse in the time-frequency plane. With impact component Among them, harmonic components Characterization Periodic acoustic characteristics, impact component Characterization The transient acoustic characteristics in the signal. Similarly, the denoised signal can be obtained. Corresponding impact component and denoised signals Corresponding impact component .

[0037] After obtaining the impulse components in each denoised signal, the system can calculate the impulse energy factor corresponding to each denoised signal based on these impulse components. The impulse energy factor characterizes the energy intensity of the impulse component. Specifically, the system can calculate the energy of each denoised signal and the energy of its corresponding impulse component, and then generate the impulse energy factor for each denoised signal based on the ratio of the energy of the impulse component to the energy of the corresponding denoised signal. For example, its corresponding impact energy factor It can be calculated using the following formula:

[0038] in, Indicates the impact component energy, Represents the denoised signal Energy.

[0039] Similarly, the denoised signal can be obtained. Corresponding impact energy factor and denoised signals Corresponding impact energy factor .

[0040] S103: For any target denoised signal among the denoised signals, calculate the outlier percentage of the impact energy factor corresponding to the target denoised signal, and evaluate the connection status of the blade corresponding to the target denoised signal based on the outlier percentage.

[0041] Since the impact energy factor can characterize the energy intensity of the impact component, it can be used as an indicator to quantify the intensity of transient anomalies and assess the connection status of wind turbine blades. For example, under normal connection conditions, the impact energy factor should remain at a low and stable level. If the impact energy factor value exceeds a preset threshold within a certain monitoring period, it indicates a significant transient impact event, possibly corresponding to bolt breakage. Furthermore, if the frequency or number of times the impact energy factor exceeds the threshold reaches a set criterion within multiple monitoring periods, it can be inferred that multiple bolts are loose.

[0042] Based on the above principles, for any target denoised signal among the various denoised signals, the system can calculate the impact energy factor sequence of the target denoised signal within a continuous monitoring period. Then, each impact energy factor in this sequence is compared with an anomaly detection threshold. If the value of any impact energy factor is greater than the threshold, it is determined to be an anomaly. By statistically analyzing the proportion of anomalies in the sequence, the frequency of transient impact events can be estimated, thereby assessing the connection status of the blade corresponding to the target denoised signal.

[0043] Because the system continuously acquires acoustic signals from each blade at a fixed period and performs denoising processing on the signals in each period, it continuously generates new multi-channel denoised signals, thus forming a denoised signal sequence arranged in chronological order. To reduce misjudgments caused by environmental disturbances of individual blades, in one embodiment, after generating the impact energy factor corresponding to each denoised signal, the system can calculate the impact energy factor corresponding to each of the periodically acquired multi-channel denoised signals, and sort all the impact energy factors corresponding to each denoised signal in chronological order to form a global impact energy factor sequence. Then, an anomaly judgment threshold is generated based on the above global impact energy factor sequence.

[0044] When generating anomaly detection thresholds based on a global impact energy factor sequence, the system first maintains a sliding time window within the global impact energy factor sequence, and then calculates distribution characteristics based on all impact energy factors within the sliding time window. Subsequently, the system generates anomaly detection thresholds based on these distribution characteristics, and periodically updates the anomaly detection thresholds when new impact energy factors are added to the sliding time window.

[0045] For example, the system can take the impact energy factor from the most recent 7 days. ={ , , ,..., , , Then, calculate its distribution characteristics (mean μ and variance) according to the following formulas. ):

[0046]

[0047] Then calculate the anomaly detection threshold according to the following formula. :

[0048] Over time, after generating new impact energy factors, the system can substitute these newly generated impact energy factors into the above formula, thereby periodically adjusting the anomaly detection threshold. Update.

[0049] In one implementation, after the system generates an anomaly detection threshold, it can calculate the proportion of outliers in the impact energy factors corresponding to the target denoised signal. Specifically, the system first obtains the target impact energy factor sequence of the target denoised signal within a preset time period, and then calculates the proportion of impact energy factors in the target impact energy factor sequence that are greater than the anomaly detection threshold, thereby generating the outlier proportion.

[0050] Continue with denoised signal For example, the system calculates the anomaly detection threshold. Then, the denoised signal can be obtained. The corresponding recent One impact energy factor { , ,..., Then, calculate the denoised signal according to the following formula. The corresponding percentage of outliers :

[0051] Similarly, the denoised signal can be obtained. The corresponding percentage of outliers and denoised signals The corresponding percentage of outliers Clearly, the proportion of outliers... Corresponding to blade A, the percentage of outliers Corresponding to blade B, the percentage of outliers Corresponds to blade C.

[0052] In one implementation, after calculating the proportion of outliers for each blade, the system can assess the risk level of the connection status of each blade according to preset rules, and send different levels of early warning signals to the alarm system or maintenance personnel based on the risk level.

[0053] For example, the system can refer to the rules in Table 1 to assess the warning level and provide handling suggestions.

[0054] Table 1 Warning Levels and Handling Recommendations Judgment conditions Warning Level Recommendations for handling Outlier percentage <20% normal none Outlier percentage ≤ 20% < 40% Level 1 alarm Perform maintenance as needed and continuously monitor trend changes. Outlier percentage ≤ 40% < 60% Level 2 alarm Timely inspection and maintenance, with a focus on trend changes. Outliers account for ≥60% Level 3 alarm Immediately inspect and repair the equipment, and take safety measures such as limiting power or shutting it down. like Figure 2 As shown, taking the impact energy factor trend chart of a certain wind turbine as an example, =30.83%, triggering the system's level one alarm, and , The result was 0 for all values, indicating that the bolts at the blade root of blade A may have become loose or broken. Ultimately, on-site maintenance personnel discovered that two bolts at the blade root of blade A had become loose and fallen off.

[0055] Please see Figure 3 This application also provides a blade connection status assessment device, which is applied to a wind turbine generator having multiple blades. The blade connection status assessment device includes: The signal acquisition module is used to periodically and synchronously acquire multiple acoustic signals from the multiple blades to extract multiple denoised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade. The signal processing module is used to extract the impulse components from each of the denoised signals to calculate the impulse energy factor corresponding to each of the denoised signals, wherein the impulse energy factor characterizes the energy intensity of the impulse component. The blade evaluation module is used to calculate the proportion of outliers of the impact energy factor corresponding to any one of the target denoised signals, and evaluate the connection status of the blade corresponding to the target denoised signal based on the proportion of outliers.

[0056] Please see Figure 4 This application also provides a blade connection status evaluation device, which includes a memory and a processor. The memory stores a computer program, and when the computer program is executed by the processor, the blade connection status evaluation method described above can be implemented. Specifically, at the hardware level, the blade connection status evaluation device may include a processor, an internal bus, and a memory. The memory may include main memory and non-volatile memory. The processor reads the corresponding computer program from the non-volatile memory into main memory and then runs it. Those skilled in the art will understand that... Figure 4 The structure shown is for illustrative purposes only and does not limit the structure of the blade connection status assessment device described above. For example, the blade connection status assessment device may also include a... Figure 4 The components shown may include more or fewer components, such as other processing hardware like a GPU (Graphics Processing Unit) or external communication ports. Of course, this application does not exclude other implementation methods besides software implementations, such as logic devices or a combination of hardware and software.

[0057] In this embodiment, the processor may include a central processing unit (CPU) or a graphics processing unit (GPU), and may also include other microcontrollers, logic gates, integrated circuits, or appropriate combinations thereof with logic processing capabilities. The memory described in this embodiment can be a storage device for storing information. In digital systems, a device capable of storing binary data can be a memory; in integrated circuits, a circuit without physical form but with storage function can also be a memory, such as RAM or FIFO; in a system, a storage device with physical form can also be called a memory. In implementation, this memory can also be implemented using a cloud storage method; the specific implementation method is not limited in this specification.

[0058] It should be noted that the specific implementation method of the blade connection status evaluation device in this specification can be referred to the description of the method implementation method, and will not be repeated here.

[0059] This application also provides a computer-readable medium storing instructions that, when executed by a processor, enable the processor to implement the blade connection status evaluation method described in the above embodiments.

[0060] This application also provides a computer program product, which includes computer program code. When the computer program code is run on a computer, the computer can implement the blade connection status evaluation method in the above embodiments.

[0061] This application also provides a chip, the chip including a circuit, the circuit being used to perform the blade connection status evaluation method in the above embodiments.

[0062] Therefore, the technical solution provided in this application first periodically and synchronously acquires multiple acoustic signals from multiple blades of a wind turbine. Then, it utilizes the structural symmetry of the multiple blades to estimate and suppress common-mode acoustic background, thereby converting each acoustic signal into a corresponding denoised signal. Next, leveraging the sparsity of structural transient anomalies in the time-frequency domain, it extracts the impact component characterizing transient acoustic features from the denoised signal to calculate the corresponding impact energy factor. Based on a sliding time window, it fuses and models all impact energy factors to generate a dynamically updated anomaly detection threshold. Finally, by statistically analyzing the proportion of anomaly values ​​in any acoustic signal's impact energy factor, it achieves a quantitative assessment of the blade root connection status. This solution, through common-mode background estimation and suppression of multi-channel acoustic signals, can effectively distinguish between local structural anomalies and environmental interference in high-noise environments using only acoustic sensors. This significantly improves the reliability of bolt loosening or breakage fault identification and allows for monitoring of blade connection status at a lower cost.

[0063] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0064] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A blade connection state evaluation method, the method being applied to a wind power generator having a plurality of blades, characterized by, The method comprises: periodically synchronously collecting multiple acoustic signals of the multiple blades to extract multiple denoised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade; extracting an impact component in each denoised signal to calculate an impact energy factor corresponding to each denoised signal, wherein the impact energy factor represents the energy intensity of the impact component; for any target denoised signal in each denoised signal, calculating an abnormal value proportion of the impact energy factor corresponding to the target denoised signal, and evaluating the connection state of the blade corresponding to the target denoised signal based on the abnormal value proportion.

2. The method of claim 1, wherein, The periodically synchronously collecting multiple acoustic signals of the multiple blades comprises: deploying at least one acoustic sensor at the same structural position of each blade; synchronously collecting signals of each acoustic sensor based on a unified time reference to obtain time-aligned multiple acoustic signals.

3. The method of claim 2, wherein, The extracting multiple denoised signals corresponding to the multiple acoustic signals comprises: calculating a common-mode acoustic background signal of the multiple acoustic signals; removing the common-mode acoustic background signal from each acoustic signal to obtain multiple denoised signals corresponding to the multiple acoustic signals.

4. The method of claim 3, wherein, The calculating a common-mode acoustic background signal of the multiple acoustic signals comprises: obtaining sampling values corresponding to each acoustic signal at the same time; performing common-mode estimation operation on each sampling value to generate the common-mode acoustic background signal, wherein the common-mode estimation operation at least includes one of arithmetic mean, median filtering or weighted average.

5. The method of claim 4, wherein, The extracting an impact component in each denoised signal comprises: performing time-frequency transformation on each denoised signal to obtain a time-frequency spectrum corresponding to each denoised signal; based on the structural sparsity characteristics of each time-frequency spectrum, separating an impact component corresponding to each denoised signal from each time-frequency spectrum, wherein the impact component is used to represent the transient acoustic characteristics in the denoised signal.

6. The method of claim 5, wherein, The calculating an impact energy factor corresponding to each denoised signal comprises: calculating the energy of each denoised signal and the energy of the impact component corresponding to each denoised signal; generating an impact energy factor corresponding to each denoised signal according to the ratio of the energy of the impact component to the energy of the corresponding denoised signal.

7. The method of claim 6, wherein, After the generating an impact energy factor corresponding to each denoised signal, the method further comprises: calculating an impact energy factor corresponding to each denoised signal based on periodically obtained multiple denoised signals, and sorting each impact energy factor in time sequence to form a global impact energy factor sequence; generating an abnormality judgment threshold based on the global impact energy factor sequence.

8. The method of claim 7, wherein, The generating an abnormality judgment threshold based on the global impact energy factor sequence comprises: maintaining a sliding time window in the global impact energy factor sequence, and calculating a distribution feature quantity based on the impact energy factors in the sliding time window; The abnormality determination threshold is generated according to the distribution characteristic quantity, and the abnormality determination threshold is periodically updated when a new impact energy factor is added to the sliding time window.

9. The method of claim 8, wherein, The method further includes: The target impact energy factor sequence of the target de-noised signal in a preset time period is obtained; The proportion of the impact energy factors greater than the abnormality determination threshold in the target impact energy factor sequence is calculated to generate the abnormal value proportion.

10. The method according to any one of claims 1 to 9, characterized in that, After the connection state of the blade corresponding to the target de-noised signal is evaluated based on the abnormal value proportion, the method further includes: According to the risk degree of the blade connection state, a different level of early warning signal is sent.

11. A blade connection state evaluation device, the device being applied to a wind power generator having a plurality of blades, characterized by, The device includes: The signal acquisition module periodically and synchronously acquires multiple acoustic signals of the multiple blades to extract multiple de-noised signals corresponding to the multiple acoustic signals, wherein each acoustic signal corresponds to one blade. The signal processing module extracts an impact component in each de-noised signal to calculate an impact energy factor corresponding to each de-noised signal, wherein the impact energy factor represents the energy intensity of the impact component. The blade evaluation module calculates an abnormal value proportion of the impact energy factor corresponding to a target de-noised signal for any one target de-noised signal in each de-noised signal, and evaluates the connection state of the blade corresponding to the target de-noised signal based on the abnormal value proportion.

12. A blade connection state evaluation device characterized by comprising: The device includes: The memory is configured to store a computer program. The processor is configured to execute the computer program stored in the memory, so that the device executes the method in any one of claims 1 to 10.

13. A computer-readable storage medium, characterized in that, The instructions are stored on the processor, and when executed by the processor, the processor implements the method in any one of claims 1 to 10.

14. A computer program product, characterised in that, The computer program product includes computer program code, which, when executed on a computer, causes the computer to implement the method in any one of claims 1 to 10.

15. A chip, characterized by The chip includes a circuit configured to execute the method in any one of claims 1 to 10.

Images