Non-contact fan blade voiceprint monitoring device and array arrangement method thereof

Abstract

The application relates to the technical field of voiceprint monitoring, and provides a non-contact fan blade voiceprint monitoring device and an array arrangement method thereof.The device comprises a shell with a sealed structure with a protection level of IP66 and above; a microphone array is used for collecting voiceprint signals generated when a fan blade is running; a signal processing module is electrically connected with the microphone array; a communication module is electrically connected with the signal processing module, supports Ethernet, Bluetooth and WiFi communication protocols, and is used for uploading processed voiceprint data to an external server; and a power module is used for providing working power supply for each module; wherein the signal processing module is configured to: based on the geometric layout of the annular microphone array, construct a spatial filtering weight vector, perform spatial filtering processing on the collected multi-channel voiceprint signals, and generate enhanced target direction voiceprint signals. The application effectively improves the accuracy and reliability of fan blade defect monitoring.

Description

Technical Field

[0001] This invention relates to the field of acoustic signature monitoring technology, and more specifically, to a non-contact acoustic signature monitoring device for wind turbine blades and its array deployment method. Background Technology

[0002] As a key component of wind turbine units, wind turbine blades operate under harsh natural environments for extended periods, making them susceptible to defects such as cracks, pinholes, and lightning strike damage. Failure to detect and address these blade defects promptly can lead to blade breakage or even complete turbine failure, resulting in significant safety accidents and economic losses. Therefore, real-time and accurate monitoring of wind turbine blade condition is of significant engineering importance.

[0003] Currently, wind turbine blade condition monitoring mainly relies on two methods: manual inspection and vibration monitoring. Manual inspection typically involves observation through binoculars or close-up inspection by climbing the tower. This method suffers from problems such as long inspection cycles, low efficiency, strong subjectivity, and significant safety hazards, making it difficult to achieve real-time monitoring of blade condition. Vibration monitoring methods involve placing accelerometers inside the blades or nacelles to collect vibration signals during blade operation, and then analyzing the vibration characteristics to determine the blade condition. However, vibration monitoring is limited by the sensor installation location, lacks sensitivity to early micro-cracks, and the sensors themselves are prone to failure due to long-term exposure to alternating loads.

[0004] In recent years, voiceprint monitoring technology has gradually gained attention as a non-contact method for monitoring equipment status. Existing voiceprint monitoring solutions typically use a single microphone to collect the operating sound of the fan and then use time-domain, frequency-domain analysis, or traditional machine learning methods for fault identification. However, these solutions suffer from the following technical problems in practical applications: First, the operating environment of wind turbine blades is complex, with multiple strong interference sources such as wind noise, gearbox noise, and generator noise. The signal-to-noise ratio of the acoustic fingerprint signal collected by a single microphone is low, making it difficult to effectively extract the acoustic fingerprint characteristics of the blade itself, resulting in low monitoring accuracy and high false alarm rate.

[0005] Secondly, existing acoustic signature acquisition devices typically use a drilling installation method, which not only damages the tower structure but also makes it difficult to adjust once the installation position is fixed, resulting in poor adaptability. Furthermore, the directionality of a single device is fixed, making it impossible to adaptively configure it based on differences in wind turbine model, blade length, and installation height.

[0006] Third, existing monitoring solutions mostly use single-point acquisition methods, which can only determine whether there is an anomaly, but cannot spatially locate the source of the crack sound, which is not conducive to maintenance personnel quickly locating the fault location.

[0007] Fourth, existing voiceprint recognition algorithms are highly dependent on positive and negative sample data, making it difficult to establish accurate diagnostic models when crack samples are scarce. Furthermore, they lack the ability to self-calibrate against installation errors of the acquisition device and changes in environmental factors, resulting in insufficient stability and reliability of the system in long-term operation.

[0008] Therefore, there is an urgent need in this field for a non-contact wind turbine blade acoustic signature monitoring device and its array deployment method that can overcome the above-mentioned defects. Through hardware structure optimization, multi-device collaborative deployment and fusion algorithm design, the acoustic signature acquisition quality can be improved, crack sound source localization can be achieved, and the system's adaptive capability can be enhanced, thereby effectively improving the accuracy and reliability of wind turbine blade defect monitoring. Summary of the Invention

[0009] The present invention provides a non-contact wind turbine blade acoustic signature monitoring device and its array deployment method, which can overcome some or all defects of the prior art.

[0010] A non-contact wind turbine blade acoustic signature monitoring device according to the present invention comprises: The housing has a sealed structure with a protection rating of IP66 or higher. A microphone array, located inside the housing, consists of at least eight digital microphones arranged in a circular array, used to collect acoustic signals generated during the operation of the wind turbine blades; The signal processing module, electrically connected to the microphone array, includes a processor with four or more cores and a main frequency of ≥1.5GHz, a ROM memory with a capacity of ≥64GB, and a RAM memory with a capacity of ≥1GB; The communication module is electrically connected to the signal processing module and supports Ethernet, Bluetooth and WiFi communication protocols. It is used to upload the processed voiceprint data to an external server. The power supply module is electrically connected to the signal processing module and the communication module, and supports DC12V power supply and POE power supply, and is used to provide working power for each module; The signal processing module is configured to: construct a spatial filtering weight vector based on the geometric layout of the ring microphone array, perform spatial filtering processing on the acquired multi-channel voiceprint signals, and generate an enhanced target direction voiceprint signal.

[0011] Preferably, the spatial filtering process includes a beamforming algorithm, the output signal of which Represented as: ; In the formula, For the number of microphones and , For the first Time-domain signals collected by one microphone, To be based on the target direction The calculated delay compensation value, The weighting coefficients are related to the geometric distribution of the ring array, and these weighting coefficients satisfy the array response in the target direction. Constraints for maximizing in the upper direction and minimizing in the non-target direction: ; In the formula, The noise covariance matrix is... The guide vector is the direction of the target.

[0012] Preferably, the signal processing module is further configured to: construct a virtual microphone channel based on the symmetrical geometric characteristics of the ring microphone array, and enhance the low-frequency acoustic signal using differential array technology, wherein the output of the differential array... Represented as: ; In the formula, These are the differential weighting coefficients. For symmetric compensation coefficients, This is the preset differential delay.

[0013] This invention provides an array deployment method for a non-contact wind turbine blade acoustic signature monitoring device, which is based on the aforementioned non-contact wind turbine blade acoustic signature monitoring device and includes the following steps: Step A: Determine the installation location on the outer wall of the wind turbine tower or at a preset distance from the tower, wherein the installation location is 3m to 4m above the ground; Step B: Install the non-contact wind turbine blade acoustic signature monitoring device at the installation location using bolt fixing or single independent rod installation methods, and no drilling is required on the tower surface during the installation process; Step C: Evenly distribute at least two of the aforementioned non-contact wind turbine blade acoustic signature monitoring devices along the circumferential direction of the wind turbine tower, with the circumferential angle between adjacent devices being... satisfy: ; In the formula, This represents the total number of devices deployed. ; Step D: Connect each of the non-contact wind turbine blade acoustic fingerprint monitoring devices to the POE optoelectronic switch at the bottom of the tower via optical fiber or network cable, and connect to the wind farm ring network through the POE optoelectronic switch to transmit the acoustic fingerprint data to the internal network server. Step E: Based on the geometric positional relationship of each deployed device, establish a multi-device collaborative positioning model to locate the sound source of the blade crack in three-dimensional space.

[0014] Preferably, the multi-device collaborative positioning model in step E includes: using the installation position of each non-contact wind turbine blade acoustic fingerprint monitoring device as a spatial reference point, constructing a hyperboloid positioning equation set by calculating the time difference of the acoustic fingerprint signal arriving at each device, and solving for the spatial coordinates of the crack sound source. The time difference positioning model is represented as follows: ; In the formula, The voiceprint signal reaches the first The device and the first The time of each device The first The device and the first The spatial coordinates of the device Speed ​​of sound; The location of the sound source is obtained by solving the following minimization problem: ; In the formula, This represents the total number of devices deployed.

[0015] Preferably, step F is also included: based on the annular array structure of the non-contact wind turbine blade acoustic signature monitoring device and the geometric relationship of the uniform arrangement of multiple devices in the circumference of the tower, a self-calibration model is constructed to correct the sound velocity deviation and device installation error caused by changes in ambient temperature and wind speed in real time. The self-calibration model establishes an error cost function using known geometric constraints. ; In the formula, For the first The device and the first The known installation distance between the devices For the first The installation position error correction amount of each device This is the correction factor for the speed of sound. For standard speed of sound, This is the regularization coefficient.

[0016] As a preferred embodiment, the single-pole installation method in step B includes: setting up independent poles on the ground or foundation around the tower, fixing the non-contact wind turbine blade acoustic signature monitoring device to the top of the pole, and ensuring that the device's sound pickup height is consistent with the installation height of the outer wall of the tower.

[0017] Preferably, the POE optoelectronic switch in step D has the following parameter characteristics: It has ≥8 optical ports, an IP40 or higher protection rating, an operating temperature range of -40℃ to +85℃, anti-static and industrial surge protection functions, and supports wide voltage power input and reverse connection protection design.

[0018] Preferably, step G is also included: receiving remote configuration commands through the communication module of the non-contact wind turbine blade acoustic signature monitoring device, and remotely adjusting the sampling rate, bit depth, directivity and gain parameters of the microphone array, wherein the sampling rate supports 48kHz or 192kHz, and the bit depth supports 16bit or 32bit.

[0019] Preferably, the method further includes step H: based on the multiple devices uniformly deployed in step C, a ring array fusion processing model is constructed, and the voiceprint signals collected by each device are weighted and fused to obtain the fused global voiceprint features. : ; In the formula, For the first The output signal of the device after spatial filtering. For feature extraction function, For the fusion weighting coefficients related to the device location, satisfy Furthermore, the weighting coefficients are adaptively adjusted based on the relative azimuth angle between each device and the blade rotation plane.

[0020] The beneficial effects of this invention are as follows: This invention employs a circular array consisting of at least eight digital microphones, combined with a beamforming algorithm for spatial filtering. By constructing a spatial filtering weight vector, the array response is maximized in the target direction (the direction of the blade rotation plane) and minimized in non-target directions (directions of interference sources such as wind noise and mechanical noise), thereby effectively suppressing background noise and significantly improving the signal-to-noise ratio of the acquired acoustic signature signal. Compared to existing technologies using a single microphone or a simple array, this invention can still extract high-resolution blade acoustic signature features in complex wind field environments, laying a reliable data foundation for subsequent fault identification.

[0021] This invention enables three-dimensional spatial localization of blade crack sound sources by uniformly deploying at least two non-contact acoustic signature monitoring devices around the circumference of the wind turbine tower and constructing a hyperboloidal localization equation system based on the time difference of acoustic signature signals arriving at each device. Maintenance personnel can quickly pinpoint the specific blade and location of the crack based on the localization results, significantly reducing troubleshooting time and improving maintenance efficiency. This solves the technical problem in existing technologies that can only identify anomalies but cannot locate the fault source.

[0022] This invention utilizes the geometric relationship of multiple devices evenly distributed around the circumference of the tower to construct a self-calibration model, establish an error cost function, and correct in real time for sound velocity deviations caused by changes in ambient temperature and wind speed, as well as device installation position errors. This self-calibration mechanism enables the system to maintain monitoring accuracy during long-term operation, overcoming the problems of system drift and decreased reliability caused by environmental changes and the accumulation of installation errors in existing technologies.

[0023] This invention employs a ring array fusion processing model to weight and fuse the acoustic signature signals collected by each device. The fusion weight coefficients are adaptively adjusted based on the relative azimuth angle between each device and the blade's rotation plane. This collaborative fusion method fully utilizes the information redundancy from multi-angle acquisition, effectively compensating for information deficiencies from single-view acquisition. Compared to existing single-point monitoring schemes, it can more comprehensively capture blade acoustic signature characteristics, improving the accuracy of crack diagnosis and its anti-interference capability.

[0024] This invention leverages the symmetrical geometry of a ring microphone array and employs differential array technology to construct a virtual microphone channel. By differentially processing the acquired signals, it effectively enhances low-frequency acoustic signature signals. Acoustic signature signals generated by early-stage blade cracks are typically dominated by low-frequency components. This technique significantly improves the system's sensitivity to detecting early-stage micro-cracks, enabling early warning of crack defects.

[0025] This invention employs bolt-fixed installation or single-pole installation, eliminating the need for drilling holes in the tower surface and avoiding structural damage. The installation height is 3m to 4m, facilitating daily maintenance while ensuring excellent acoustic signature collection. Furthermore, the device supports PoE power supply and daisy-chain deployment, simplifying wiring and facilitating construction. The device's directional capability is configurable for omnidirectional, 90°, and 120° angles, allowing for flexible adjustment based on wind turbine model, blade length, and installation scenario, thus offering broad applicability.

[0026] This invention receives remote configuration commands through a communication module, enabling remote adjustment of sampling rate (48kHz / 192kHz selectable), bit depth (16bit / 32bit selectable), directivity, and gain parameters without on-site operation, significantly reducing maintenance costs and improving system maintainability. Attached Figure Description

[0027] Figure 1 This is a structural block diagram of a non-contact wind turbine blade acoustic signature monitoring device in one embodiment. Detailed Implementation

[0028] To further understand the content of this invention, a detailed description of the invention will be provided in conjunction with the accompanying drawings and embodiments. It should be understood that the embodiments are merely illustrative and not limiting of the invention. Example

[0029] like Figure 1 As shown, this embodiment provides a non-contact wind turbine blade acoustic signature monitoring device, characterized in that it includes: The housing has a sealed structure with a protection rating of IP66 or higher. A microphone array, located inside the housing, consists of at least eight digital microphones arranged in a circular array, used to collect acoustic signals generated during the operation of the wind turbine blades; The signal processing module, electrically connected to the microphone array, includes a processor with four or more cores and a main frequency of ≥1.5GHz, a ROM memory with a capacity of ≥64GB, and a RAM memory with a capacity of ≥1GB; The communication module is electrically connected to the signal processing module and supports Ethernet, Bluetooth and WiFi communication protocols. It is used to upload the processed voiceprint data to an external server. The power supply module is electrically connected to the signal processing module and the communication module, and supports DC12V power supply and POE power supply, and is used to provide working power for each module; The signal processing module is configured to: construct a spatial filtering weight vector based on the geometric layout of the ring microphone array, perform spatial filtering processing on the acquired multi-channel voiceprint signals, and generate an enhanced target direction voiceprint signal.

[0030] In this embodiment, the spatial filtering process includes a beamforming algorithm, whose output signal Represented as: ; In the formula, For the number of microphones and , For the first Time-domain signals collected by one microphone, To be based on the target direction The calculated delay compensation value, The weighting coefficients are related to the geometric distribution of the ring array, and these weighting coefficients satisfy the array response in the target direction. Constraints for maximizing in the upper direction and minimizing in the non-target direction: ; In the formula, The noise covariance matrix is... The guide vector is the direction of the target.

[0031] In this embodiment, the signal processing module is further configured to: construct a virtual microphone channel based on the symmetrical geometric characteristics of the ring microphone array, and enhance the low-frequency acoustic signature signal using differential array technology, wherein the output of the differential array... Represented as:

[0032] In the formula, These are the differential weighting coefficients. For symmetric compensation coefficients, This is the preset differential delay.

[0033] This embodiment provides an array deployment method for a non-contact wind turbine blade acoustic signature monitoring device, which is based on the aforementioned non-contact wind turbine blade acoustic signature monitoring device and includes the following steps: Step A: Determine the installation location on the outer wall of the wind turbine tower or at a preset distance from the tower, wherein the installation location is 3m to 4m above the ground; Step B: Install the non-contact wind turbine blade acoustic signature monitoring device at the installation location using bolt fixing or single independent rod installation methods, and no drilling is required on the tower surface during the installation process; Step C: Evenly distribute at least two of the aforementioned non-contact wind turbine blade acoustic signature monitoring devices along the circumferential direction of the wind turbine tower, with the circumferential angle between adjacent devices being... satisfy: ; In the formula, This represents the total number of devices deployed. ; Step D: Connect each of the non-contact wind turbine blade acoustic fingerprint monitoring devices to the POE optoelectronic switch at the bottom of the tower via optical fiber or network cable, and connect to the wind farm ring network through the POE optoelectronic switch to transmit the acoustic fingerprint data to the internal network server. Step E: Based on the geometric positional relationship of each deployed device, establish a multi-device collaborative positioning model to locate the sound source of the blade crack in three-dimensional space.

[0034] In this embodiment, the multi-device collaborative positioning model in step E includes: using the installation position of each non-contact wind turbine blade acoustic fingerprint monitoring device as a spatial reference point, constructing a hyperboloid positioning equation set by calculating the time difference of the acoustic fingerprint signal arriving at each device, and solving for the spatial coordinates of the crack sound source. The time difference positioning model is represented as follows:

[0035] In the formula, The voiceprint signal reaches the first The device and the first The time of each device The first The device and the first The spatial coordinates of the device Speed ​​of sound; The location of the sound source is obtained by solving the following minimization problem: ; In the formula, This represents the total number of devices deployed.

[0036] In this embodiment, step F is also included: based on the annular array structure of the non-contact wind turbine blade acoustic signature monitoring device and the geometric relationship of the uniform arrangement of multiple devices around the tower, a self-calibration model is constructed to correct the sound velocity deviation and device installation error caused by changes in ambient temperature and wind speed in real time. The self-calibration model establishes an error cost function using known geometric constraints. ; In the formula, For the first The device and the first The known installation distance between the devices For the first The installation position error correction amount of each device This is the correction factor for the speed of sound. For standard speed of sound, This is the regularization coefficient.

[0037] In this embodiment, the single pole installation method in step B includes: setting up independent poles on the ground or foundation around the tower, fixing the non-contact wind turbine blade acoustic signature monitoring device to the top of the pole, and keeping the device's sound pickup height consistent with the installation height of the outer wall of the tower.

[0038] In this embodiment, the POE optoelectronic switch in step D has the following parameter characteristics: It has ≥8 optical ports, an IP40 or higher protection rating, an operating temperature range of -40℃ to +85℃, anti-static and industrial surge protection functions, and supports wide voltage power input and reverse connection protection design.

[0039] In this embodiment, step G is also included: receiving remote configuration instructions through the communication module of the non-contact wind turbine blade acoustic print monitoring device, and remotely adjusting the sampling rate, bit depth, directivity and gain parameters of the microphone array, wherein the sampling rate supports 48kHz or 192kHz and the bit depth supports 16bit or 32bit.

[0040] In this embodiment, step H is also included: based on the multiple devices uniformly deployed in step C, a ring array fusion processing model is constructed, and the voiceprint signals collected by each device are weighted and fused to obtain the fused global voiceprint features. : ; In the formula, For the first The output signal of the device after spatial filtering. For feature extraction function, For the fusion weighting coefficients related to the device location, satisfy Furthermore, the weighting coefficients are adaptively adjusted based on the relative azimuth angle between each device and the blade rotation plane.

[0041] This embodiment employs a circular array consisting of at least eight digital microphones, combined with a beamforming algorithm for spatial filtering. By constructing a spatial filtering weight vector, the array response is maximized in the target direction (blade rotation plane) and minimized in non-target directions (directions of interference sources such as wind noise and mechanical noise), thereby effectively suppressing background noise and significantly improving the signal-to-noise ratio of the acquired acoustic signature signal. Compared to existing technologies using a single microphone or a simple array, this invention can still extract high-resolution blade acoustic signature features in complex wind field environments, laying a reliable data foundation for subsequent fault identification.

[0042] This embodiment utilizes at least two non-contact acoustic signature monitoring devices evenly distributed around the circumference of the wind turbine tower. Based on the time difference of the acoustic signature signals arriving at each device, a hyperboloidal localization equation system is constructed, enabling three-dimensional spatial localization of the sound source of blade cracks. Maintenance personnel can quickly pinpoint the specific blade and location of the crack based on the localization results, significantly reducing troubleshooting time and improving maintenance efficiency. This solves the technical problem in existing technologies that can only identify anomalies but cannot locate the fault source.

[0043] This embodiment utilizes the geometric relationship of multiple devices evenly distributed around the tower circumference to construct a self-calibration model, establish an error cost function, and correct in real time for sound velocity deviations caused by changes in ambient temperature and wind speed, as well as device installation position errors. This self-calibration mechanism enables the system to maintain monitoring accuracy during long-term operation, overcoming the problems of system drift and decreased reliability caused by environmental changes and the accumulation of installation errors in existing technologies.

[0044] This embodiment employs a ring array fusion processing model to weight and fuse the acoustic signature signals collected by each device. The fusion weight coefficients are adaptively adjusted based on the relative azimuth angle between each device and the blade's rotation plane. This collaborative fusion method fully utilizes the information redundancy from multi-angle acquisition, effectively compensating for the information loss from single-view acquisition. Compared to existing single-point monitoring schemes, it can more comprehensively capture blade acoustic signature characteristics, improving the accuracy of crack diagnosis and its anti-interference capability.

[0045] This embodiment leverages the symmetrical geometry of a ring microphone array and employs differential array technology to construct a virtual microphone channel. Differential processing of the acquired signals effectively enhances low-frequency acoustic signature signals. Acoustic signature signals generated by early-stage blade cracks are typically dominated by low-frequency components. This technique significantly improves the system's sensitivity to detecting early-stage micro-cracks, enabling early warning of crack defects.

[0046] This embodiment employs bolt-fixed installation or single-pole installation, eliminating the need for drilling holes in the tower surface and avoiding structural damage. The installation height is 3m to 4m, facilitating daily maintenance while ensuring good acoustic signature collection. Simultaneously, the device supports PoE power supply and daisy-chain deployment, simplifying wiring and facilitating construction. The device's directivity is configurable in omnidirectional, 90°, and 120° configurations, allowing for flexible adjustment based on wind turbine model, blade length, and installation scenario, thus offering wide applicability.

[0047] This embodiment receives remote configuration commands through the communication module, which can remotely adjust the sampling rate (48kHz / 192kHz selectable), bit depth (16bit / 32bit selectable), directivity and gain parameters without on-site operation, greatly reducing operation and maintenance costs and improving system maintainability.

[0048] The present invention and its embodiments have been described above illustratively. This description is not restrictive, and the figures shown are only one embodiment of the present invention; the actual structure is not limited thereto. Therefore, if those skilled in the art are inspired by this description and design similar structures and embodiments without departing from the spirit of the present invention, such designs should fall within the protection scope of the present invention.

Claims

1. A non-contact acoustic signature monitoring device for wind turbine blades, characterized in that, include: The housing has a sealed structure with a protection rating of IP66 or higher. A microphone array, located inside the housing, consists of at least eight digital microphones arranged in a circular array, used to collect acoustic signals generated during the operation of the wind turbine blades; The signal processing module, electrically connected to the microphone array, includes a processor with four or more cores and a main frequency of ≥1.5GHz, a ROM memory with a capacity of ≥64GB, and a RAM memory with a capacity of ≥1GB; The communication module is electrically connected to the signal processing module and supports Ethernet, Bluetooth and WiFi communication protocols. It is used to upload the processed voiceprint data to an external server. The power supply module is electrically connected to the signal processing module and the communication module, and supports DC12V power supply and POE power supply, and is used to provide working power for each module; The signal processing module is configured to: construct a spatial filtering weight vector based on the geometric layout of the ring microphone array, perform spatial filtering processing on the acquired multi-channel voiceprint signals, and generate an enhanced target direction voiceprint signal.

2. The non-contact wind turbine blade acoustic signature monitoring device according to claim 1, characterized in that, The spatial filtering process includes a beamforming algorithm, whose output signal Represented as: ； In the formula, For the number of microphones and , For the first Time-domain signals collected by one microphone, To be based on the target direction The calculated delay compensation value, The weighting coefficients are related to the geometric distribution of the ring array, and these weighting coefficients satisfy the constraint that the array response is maximized in the target direction and minimized in the non-target direction. ; In the formula, The noise covariance matrix is... The guide vector is the direction of the target.

3. The non-contact wind turbine blade acoustic signature monitoring device according to claim 1, characterized in that, The signal processing module is further configured to: construct a virtual microphone channel based on the symmetrical geometric characteristics of the ring microphone array, and enhance the low-frequency acoustic signal using differential array technology, wherein the output of the differential array... Represented as: ； In the formula, These are the differential weighting coefficients. For symmetric compensation coefficients, This is the preset differential delay.

4. A method for arraying a non-contact wind turbine blade acoustic signature monitoring device, characterized in that, It is based on a non-contact wind turbine blade acoustic signature monitoring device according to any one of claims 1-3, and includes the following steps: Step A: Determine the installation location on the outer wall of the wind turbine tower or at a preset distance from the tower, wherein the installation location is 3m to 4m above the ground; Step B: Install the non-contact wind turbine blade acoustic signature monitoring device at the installation location using bolt fixing or single independent rod installation methods, and no drilling is required on the tower surface during the installation process; Step C: At least two of the aforementioned non-contact wind turbine blade acoustic signature monitoring devices are evenly distributed along the circumferential direction of the wind turbine tower, with the circumferential angle between adjacent devices satisfying: ； In the formula, This represents the total number of devices deployed. ; Step D: Connect each of the non-contact wind turbine blade acoustic fingerprint monitoring devices to the POE optoelectronic switch at the bottom of the tower via optical fiber or network cable, and connect to the wind farm ring network through the POE optoelectronic switch to transmit the acoustic fingerprint data to the internal network server. Step E: Based on the geometric positional relationship of each deployed device, establish a multi-device collaborative positioning model to locate the sound source of the blade crack in three-dimensional space.

5. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, The multi-device cooperative localization model mentioned in step E includes: Using the installation locations of the acoustic fingerprint monitoring devices for each non-contact wind turbine blade as spatial reference points, the spatial coordinates of the crack sound source are solved by calculating the time difference between the arrival of the acoustic fingerprint signal at each device and constructing a hyperboloidal positioning equation system. The time difference positioning model is represented as follows: ; In the formula, The voiceprint signal reaches the first The device and the first The time of each device The first The device and the first The spatial coordinates of the device Speed ​​of sound; The location of the sound source is obtained by solving the following minimization problem: ； In the formula, This represents the total number of devices deployed.

6. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, The method also includes step F: based on the annular array structure of the non-contact wind turbine blade acoustic signature monitoring device and the geometric relationship of the uniform arrangement of multiple devices around the tower, a self-calibration model is constructed to correct the sound velocity deviation and device installation error caused by changes in ambient temperature and wind speed in real time. The self-calibration model establishes an error cost function using known geometric constraints. ； In the formula, For the first The device and the first Known installation distances between devices For the first The installation position error correction amount of each device This is the correction factor for the speed of sound. For standard speed of sound, is the regularization coefficient.

7. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, The single-pole installation method described in step B includes: setting up independent poles on the ground or foundation around the tower, fixing the non-contact wind turbine blade acoustic signature monitoring device to the top of the pole, and ensuring that the device's sound pickup height is consistent with the installation height of the outer wall of the tower.

8. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, The POE optoelectronic switch described in step D has the following parameter characteristics: optical port It has ≥8 ports, an IP40 or higher protection rating, an operating temperature range of -40℃ to +85℃, anti-static and industrial surge protection functions, and supports wide voltage power input and reverse connection protection design.

9. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, It also includes step G: receiving remote configuration commands through the communication module of the non-contact wind turbine blade acoustic signature monitoring device, and remotely adjusting the sampling rate, bit depth, directivity and gain parameters of the microphone array, wherein the sampling rate supports 48kHz or 192kHz, and the bit depth supports 16bit or 32bit.

10. The array deployment method of a non-contact wind turbine blade acoustic signature monitoring device according to claim 4, characterized in that, The process also includes step H: Based on the multiple devices uniformly deployed in step C, a ring array fusion processing model is constructed, and the voiceprint signals collected by each device are weighted and fused to obtain the fused global voiceprint features. : ； In the formula, For the first The output signal of the device after spatial filtering. For feature extraction function, For the fusion weighting coefficients related to the device location, satisfy Furthermore, the weighting coefficients are adaptively adjusted based on the relative azimuth angle between each device and the blade rotation plane.

Images