Wind turbine generator voiceprint self-adaptive detection method, system and equipment and storage medium

Abstract

The invention discloses a wind turbine generator voiceprint self-adaptive detection method, system and device and a storage medium, and belongs to the technical field of wind power, and the method comprises the steps: collecting a voiceprint signal, raindrop diameter, environment humidity, environment temperature and atmospheric pressure of a wind turbine generator in real time through a sensor array; temperature and humidity coupling correction parameters are calculated based on the environment humidity, and real-time air density is obtained through correction according to an air density correction formula by means of the environment humidity, the environment temperature and the atmospheric pressure; constructing a rain intensity noise spectrum mapping model based on the raindrop diameter and the real-time air density in combination with a raindrop vertical fragmentation merging effect compensation algorithm; performing rain noise suppression on the voiceprint signal of the wind turbine generator by adopting a depth spectrum subtraction algorithm in combination with the rain intensity noise spectrum mapping model and the dynamic suppression factor to obtain a clean voiceprint signal; and matching the clean voiceprint signal with a preset wind turbine generator fault voiceprint feature library to complete wind turbine generator voiceprint adaptive detection.

Description

Technical Field

[0001] This invention belongs to the field of wind power technology, specifically relating to the adaptive detection method, system, equipment, and storage medium for wind turbine acoustic signatures. Background Technology

[0002] In scenarios such as industrial equipment condition monitoring and outdoor acoustic testing, effective suppression of environmental noise is crucial for ensuring the quality of sound data and the accuracy of subsequent feature analysis. Rain noise generated by rainfall is a typical and difficult-to-handle source of interference. Currently, the industry commonly uses traditional bandpass filtering technology to suppress rain noise. This technology filters out sound signals in the target frequency band by setting a fixed cutoff frequency to eliminate rain noise interference, but it has significant limitations in practical applications.

[0003] The energy distribution of rain noise changes dynamically with rainfall intensity. For example, during light rain, rain noise energy is mainly concentrated in the lower frequency band, while during heavy rain, the energy extends to the mid-to-high frequency band. The fixed cutoff frequency of traditional bandpass filters cannot adapt to this dynamic characteristic. In heavy rain scenarios, fixed filter frequency bands can easily lead to spectral confusion between rain noise and the target sound signal. This not only fails to effectively filter out rain noise but may also result in the loss of key information from the target signal, leading to equipment malfunctions, false alarms, and decreased detection accuracy, seriously affecting the reliability of the monitoring system.

[0004] Further analysis reveals that the core energy of rain noise is concentrated in the mid-to-low frequency band of 200-800Hz, while the frequency band of gearbox wear faults, a key concern in industrial applications, is 500-1500Hz. There is significant overlap between the two in the 500-800Hz range. Existing filtering technologies, limited by a single frequency selection mechanism, struggle to completely separate rain noise interference within the overlapping frequency bands while preserving gearbox wear characteristic signals. This results in a large amount of rain noise remaining in the audio data, failing to meet the audio data quality requirements of high-precision equipment condition monitoring. Therefore, a novel noise suppression technology is urgently needed to dynamically adapt to changes in rain noise and effectively separate interference in overlapping frequency bands for adaptive detection of wind turbine acoustic signatures. Summary of the Invention

[0005] The purpose of this invention is to overcome the problems of wind turbine acoustic signature detection being unable to dynamically adapt to changes in rain noise and effectively separate overlapping frequency band interference, and to propose an adaptive acoustic signature detection method, system, equipment and storage medium for wind turbines.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides an adaptive acoustic signature detection method for wind turbine generators, comprising the following steps: Step S1: Use a sensor array to collect real-time acoustic signature signals from the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure. Step S2: Calculate temperature and humidity coupling correction parameters based on ambient humidity, and use ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula; Step S3: Construct a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm; Step S4: Using the depth spectrum subtraction algorithm, combined with the rain intensity noise spectrum mapping model and dynamic suppression factor, rain noise suppression is performed on the wind turbine acoustic signature signal to obtain a clean acoustic signature signal. Step S5: Match the cleaning acoustic signature signal with the preset wind turbine fault acoustic signature feature library to complete the wind turbine acoustic signature adaptive detection.

[0007] Furthermore, the sensor array in step S1 includes an acoustic fingerprint sensor, an optical rain sensor, and an environmental sensor. The acoustic fingerprint sensor is installed at the root of the wind turbine blades, the gearbox housing, and the middle section of the tower. Optical rain sensors are installed on the top of the wind turbine tower and on the windward side of the blades. Environmental sensors include temperature and humidity sensors, and barometers; In step S1, the acoustic signature signal of the wind turbine is collected using an acoustic signature sensor, the raindrop diameter is collected by inversion using an optical rain sensor combined with the infrared light shading effect, and the ambient humidity, ambient temperature and atmospheric pressure are monitored and collected using an environmental sensor.

[0008] Furthermore, in step S2, the temperature and humidity coupling correction parameters are calculated based on the ambient humidity as shown in the following formula:

[0009] in, For temperature and humidity coupling correction parameters, For real-time ambient humidity, , , To correct the parameters; The air density correction formula in step S2 is shown below:

[0010] in, For real-time air density, Relative humidity, Atmospheric pressure, For ambient temperature, This is the partial pressure of saturated water vapor.

[0011] Further, step S3 constructs a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm, including the following steps: Step S301: Combine rainfall intensity and average raindrop diameter to fit and determine the parameters of the raindrop diameter distribution function, and establish the raindrop diameter distribution function using the Gamma distribution; Step S302: Calculate the power spectrum of rain noise generated by raindrop impacting wind turbine components based on the raindrop kinetic energy distribution and real-time air density; Step S303: Based on the height difference between the sensor and the cloud layer, perform vertical height compensation on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights; Step S304: Use a hybrid Copula function to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum to obtain a rainfall intensity noise spectrum mapping model.

[0012] Furthermore, in step S304, the hybrid Copula function includes at least a sub-Copula function for capturing high-frequency tail correlations and strengthening the association of low-probability events, and the function weights are optimized by Bayesian weighted average.

[0013] Furthermore, in step S301, a raindrop diameter distribution function is established using the Gamma distribution, as shown in the following equation:

[0014] in, The diameter of the raindrop. For the intercept parameter, For shape parameters, The slope parameter; In step S302, the power spectrum of rain noise generated by raindrops impacting wind turbine components is calculated based on the raindrop kinetic energy distribution and real-time air density, as shown in the following formula:

[0015]

[0016] in, The power spectrum of rain noise. The final velocity of the raindrop. The drag coefficient, For real-time air density, For the density of water, For correction factor, For frequency variables; In step S303, based on the height difference between the sensor and the cloud layer, vertical height compensation is performed on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights, as shown in the following formula:

[0017] in, The difference in height between the sensor and the cloud layer; In step S304, a hybrid Copula function is used to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum, resulting in a rainfall intensity-noise spectrum mapping model, as shown in the following equation:

[0018] in, Let be the edge distribution function of the power spectrum of rainfall intensity and rainfall noise. These are the weight parameters.

[0019] Furthermore, in step S4, the dynamic suppression factor is combined with the temperature and humidity coupling correction coefficient and the dynamic adjustment of rainfall intensity, and zero suppression is implemented in the fault characteristic frequency band of the wind turbine, while it is linearly enhanced with rainfall intensity in the main energy area of ​​rain noise. The depth spectrum subtraction algorithm in step S4 is shown in the following equation:

[0020]

[0021] in, As a dynamic inhibitor, To correct the parameters, Rain intensity.

[0022] Secondly, the present invention provides a wind turbine acoustic signature adaptive detection system, comprising: The sensor acquisition module is used to acquire real-time data such as acoustic signature of the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure using a sensor array. The correction parameter module is used to calculate temperature and humidity coupled correction parameters based on ambient humidity. It uses ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula. A mapping model module is constructed to build a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm. The rain noise suppression module is used to suppress rain noise in the acoustic signature signal of the wind turbine by employing a depth spectrum subtraction algorithm, combined with a rain intensity noise spectrum mapping model and a dynamic suppression factor, to obtain a clean acoustic signature signal. The adaptive acoustic signature detection module is used to match the clean acoustic signature signal with a preset acoustic signature feature library of wind turbine faults to complete the adaptive acoustic signature detection of wind turbines.

[0023] Thirdly, the present invention provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the wind turbine acoustic signature adaptive detection method.

[0024] Fourthly, the present invention provides a computer-readable storage medium storing a computer program, wherein the computer program, when executed by a processor, describes the wind turbine acoustic signature adaptive detection method.

[0025] Compared with the prior art, the present invention has the following beneficial technical effects: This invention proposes an adaptive acoustic signature detection method for wind turbine generators. It constructs a rain intensity noise spectrum mapping model with multimodal collaborative noise reduction, integrating acoustic, optical, and environmental data to generate a dynamic noise spectrum. It also incorporates a bidirectional mapping mechanism driven by a DSD physical model and a hybrid Copula-SVM data driver. The rain intensity noise spectrum mapping model compensates for vertical effects through a raindrop vertical breakup and merging effect compensation algorithm. Rain noise suppression uses a dynamic suppression factor to dynamically suppress and filter rainstorm signals. This invention can dynamically adapt to changes in rain noise and effectively separate overlapping frequency band interference to achieve adaptive acoustic signature detection for wind turbine generators. By collecting multi-dimensional environmental parameters such as raindrop diameter, temperature, humidity, and atmospheric pressure, this invention constructs a temperature and humidity coupled correction parameter and a real-time air density correction model to solve the interference problem of environmental factors on acoustic signature signal detection and improve detection stability in complex environments. Combining the raindrop vertical breakup and merging effect compensation algorithm with the rain intensity noise spectrum mapping model, along with a depth spectrum subtraction algorithm and a dynamic suppression factor, can specifically filter out rain noise, obtain high-quality clean acoustic signature signals, and reduce false or missed detections caused by noise. By matching clean acoustic signals with fault acoustic feature databases, the input quality is optimized from the signal source, the accuracy of fault feature matching is improved, the accuracy and timeliness of wind turbine fault detection are ensured, and a reliable basis is provided for equipment operation and maintenance. Attached Figure Description

[0026] The accompanying drawings described herein are for illustrative purposes only and are not intended to limit the scope of the invention in any way. Furthermore, the shapes and proportions of the components in the drawings are merely schematic to aid in understanding the invention and do not specifically limit the shapes and proportions of the components. In the drawings: Figure 1 This is a flowchart of the wind turbine acoustic signature adaptive detection method of the present invention.

[0027] Figure 2 This is a structural diagram of the wind turbine acoustic signature adaptive detection system of the present invention.

[0028] Figure 3 This is a diagram of the electronic device used in the wind turbine acoustic signature adaptive detection method of the present invention.

[0029] Figure 4 This is a flowchart illustrating the construction of a rain intensity noise spectrum mapping model using the method of this invention.

[0030] Figure 5This is a simplified system structure diagram of the wind turbine acoustic signature adaptive detection method in an embodiment of the present invention. Detailed Implementation

[0031] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0032] Example 1 See Figure 1 The adaptive detection method for acoustic signatures of wind turbine generators includes the following steps: Step S1: Use a sensor array to collect real-time acoustic signature signals from the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure. Step S2: Calculate temperature and humidity coupling correction parameters based on ambient humidity, and use ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula; Step S3: Construct a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm; Step S4: Using the depth spectrum subtraction algorithm, combined with the rain intensity noise spectrum mapping model and dynamic suppression factor, rain noise suppression is performed on the wind turbine acoustic signature signal to obtain a clean acoustic signature signal. Step S5: Match the cleaning acoustic signature signal with the preset wind turbine fault acoustic signature feature library to complete the wind turbine acoustic signature adaptive detection.

[0033] This embodiment utilizes a sensor array to comprehensively collect various types of data, accurately reflecting the unit's operating environment. By calculating temperature and humidity coupling correction parameters, correcting air density, and constructing a rain intensity noise spectrum mapping model, interference from environmental factors on the acoustic signature signal can be effectively compensated. A depth spectral subtraction algorithm combined with a dynamic suppression factor is used for rain noise suppression, enabling the acquisition of high-quality, clean acoustic signature signals. Finally, these signals are matched with a fault acoustic signature feature database to achieve adaptive detection, helping to promptly identify unit faults, reduce operation and maintenance costs, and ensure the stable operation of wind turbine units.

[0034] In step S1, the sensor array includes an acoustic fingerprint sensor, an optical rain sensor, and an environmental sensor. The acoustic fingerprint sensor is installed at the root of the wind turbine blades, the gearbox housing, and the middle section of the tower. The optical rain sensor is installed at the top of the wind turbine tower and the windward side of the blades. The environmental sensor includes a temperature and humidity sensor and a barometer. In step S1, the acoustic signature signal of the wind turbine is collected using an acoustic signature sensor, the raindrop diameter is collected by inversion using an optical rain sensor combined with the infrared light shading effect, and the ambient humidity, ambient temperature and atmospheric pressure are monitored and collected using an environmental sensor.

[0035] In step S2, the temperature and humidity coupling correction parameters are calculated based on the ambient humidity, as shown in the following formula:

[0036] in, For temperature and humidity coupling correction parameters, For real-time ambient humidity, , , To correct the parameters; The air density correction formula in step S2 is shown below:

[0037] in, For real-time air density, Relative humidity, Atmospheric pressure, For ambient temperature, This is the partial pressure of saturated water vapor.

[0038] Step S3 constructs a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm. See [link / reference]. Figure 4 This includes the following steps: Step S301: Combine rainfall intensity and average raindrop diameter to fit and determine the parameters of the raindrop diameter distribution function, and establish the raindrop diameter distribution function using the Gamma distribution; Step S302: Calculate the power spectrum of rain noise generated by raindrop impacting wind turbine components based on the raindrop kinetic energy distribution and real-time air density; Step S303: Based on the height difference between the sensor and the cloud layer, perform vertical height compensation on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights; Step S304: Use a hybrid Copula function to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum to obtain a rainfall intensity noise spectrum mapping model.

[0039] In step S304, the hybrid Copula function includes at least a sub-Copula function for capturing high-frequency tail correlations and strengthening the association of low-probability events, and the function weights are optimized by Bayesian weighted average.

[0040] In step S301, the raindrop diameter distribution function is established using the Gamma distribution, as shown in the following equation:

[0041] in, The diameter of the raindrop. For the intercept parameter, For shape parameters, The slope parameter; In step S302, the power spectrum of rain noise generated by raindrops impacting wind turbine components is calculated based on the raindrop kinetic energy distribution and real-time air density, as shown in the following formula:

[0042]

[0043] in, The power spectrum of rain noise. The final velocity of the raindrop. The drag coefficient, For real-time air density, For the density of water, For correction factor, For frequency variables; In step S303, based on the height difference between the sensor and the cloud layer, vertical height compensation is performed on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights, as shown in the following formula:

[0044] in, The difference in height between the sensor and the cloud layer; In step S304, a hybrid Copula function is used to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum, resulting in a rainfall intensity-noise spectrum mapping model, as shown in the following equation:

[0045] in, The marginal distribution function of rainfall intensity and noise spectrum. These are the weight parameters.

[0046] In step S4, the dynamic suppression factor is combined with the temperature and humidity coupling correction coefficient and the rainfall intensity is dynamically adjusted. Zero suppression is implemented in the fault characteristic frequency band of the wind turbine, and it is linearly enhanced with the rainfall intensity in the main energy area of ​​rain noise. The depth spectrum subtraction algorithm in step S4 is shown in the following equation:

[0047]

[0048] in, As a dynamic inhibitor, To correct the parameters, Rain intensity.

[0049] In this embodiment, the sensor array is rationally arranged for data acquisition, with acoustic fingerprint sensors placed in key locations to accurately capture the unit's acoustic fingerprints. Optical rain sensors and environmental sensors comprehensively acquire data such as raindrop diameter and ambient temperature and humidity, providing rich data for subsequent analysis. When constructing the rain intensity noise spectrum mapping model, a Gamma distribution is used to establish the raindrop diameter distribution function, which is scientifically sound. Multi-step calculation of the rain noise power spectrum and vertical height compensation accurately reflect the rain noise characteristics under different conditions. A hybrid Copula function is used to model nonlinear relationships, and weight optimization improves model accuracy. A dynamic suppression factor is dynamically adjusted based on multiple factors, achieving zero suppression in the fault frequency band and increasing rain noise with rainfall intensity in the rain noise region, effectively suppressing rain noise. The depth spectrum subtraction algorithm combined with the dynamic suppression factor yields high-quality clean acoustic fingerprint signals, which are matched with the fault acoustic fingerprint feature library, enabling accurate and adaptive unit fault detection and ensuring stable unit operation.

[0050] This embodiment achieves adaptive acoustic signature detection for wind turbines across the entire process from parameter acquisition to environmental correction, noise modeling, rain noise suppression, and fault detection. It employs multi-dimensional sensors to collaboratively acquire acoustic signatures, raindrop diameters, and environmental parameters, ensuring data integrity. Environmental parameter correction, temperature and humidity coupling, and air density calculation enhance noise modeling accuracy. A rain intensity noise spectrum mapping model is constructed based on Gamma distribution and a hybrid Copula function to accurately characterize rain noise features. A dynamic suppression factor combined with a deep spectral subtraction algorithm efficiently suppresses rain noise while preserving fault characteristics. Finally, by matching with a preset fault feature library, adaptive detection results are output, making it suitable for fault diagnosis scenarios of wind turbines in rainy environments.

[0051] Example 2 See Figure 2 The wind turbine acoustic signature adaptive detection system includes: The sensor acquisition module is used to acquire real-time data such as acoustic signature of the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure using a sensor array. The correction parameter module is used to calculate temperature and humidity coupled correction parameters based on ambient humidity. It uses ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula. A mapping model module is built to construct a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density; The rain noise suppression module is used to suppress rain noise in the acoustic signature signal of the wind turbine by employing a depth spectrum subtraction algorithm, combined with a rain intensity noise spectrum mapping model and a dynamic suppression factor, to obtain a clean acoustic signature signal. The adaptive acoustic signature detection module is used to match the clean acoustic signature signal with a preset acoustic signature feature library of wind turbine faults to complete the adaptive acoustic signature detection of wind turbines.

[0052] Example 3 See Figure 3 An electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the wind turbine acoustic signature adaptive detection method.

[0053] Example 4 A computer-readable storage medium storing a computer program, which, when executed by a processor, describes an adaptive detection method for wind turbine acoustic signatures.

[0054] Example 5 like Figure 5 As shown in the figure, the wind turbine acoustic signature adaptive detection method proposed in this embodiment is based on acoustic waveguide compensation and multimodal rain noise suppression. The main process is to establish a rain intensity-noise spectrum mapping model, suppress rain noise under multimodal conditions, and match fault acoustic signature features. The following is a detailed description of the specific implementation process of the wind turbine acoustic signature adaptive detection method: (1) Sensor array deployment: Acoustic sensors: Acoustic sensors are deployed at the blade roots, gearbox housing, and middle section of the tower; high-precision piezoelectric acoustic sensors are selected and calibrated using a standard acoustic calibrator before installation; Optical rain sensor: An optical rain sensor is deployed on the top of the wind turbine tower and the windward side of the blades to invert the diameter D of raindrops through the infrared light blocking effect; Environmental sensors: Deploy temperature and humidity sensors and barometers to monitor ambient humidity H (%), temperature T (°C), and atmospheric pressure P. Install temperature and humidity sensors and barometers on the outer wall of the control room at the bottom of the tower.

[0055] The data acquisition unit is equipped with a voiceprint signal acquisition module and an analog signal acquisition module. All sensor signals are connected to the corresponding modules via shielded cables. The host computer is an industrial control computer, which installs software for data acquisition and real-time preprocessing, and sets the acquisition frequency and data storage format.

[0056] (2) Environmental parameter correction: During unit operation, environmental sensors collect temperature and humidity (T, H) and air pressure (P) data in real time, updating them every set time interval. The host computer then corrects the environmental parameters based on the collected data. Temperature-humidity coupling correction:

[0057] Real-time ambient humidity , , To correct the parameters; Air density correction:

[0058] Atmospheric pressure (MPa). Ambient temperature (°C) Relative humidity : Saturated water vapor partial pressure; (3) Construct a rainfall intensity-noise spectrum mapping model An optical rain gauge collects raindrop diameter D and rainfall intensity data. The host computer then constructs a mapping model using the following steps: Calculating raindrop diameter using Gamma distribution (mm):

[0059] Raindrop diameter (mm) : Intercept parameter; Shape parameters (to distinguish between continental rainfall, maritime rainfall, etc.); : Slope parameter, from and (The mass-weighted average diameter) is obtained through fitting.

[0060] Noise power spectrum generated by raindrops hitting the blade surface The rain noise power spectrum is determined by the distribution of raindrop kinetic energy:

[0061] Raindrop terminal velocity :

[0062] The drag coefficient, The density of water; Correction factor, It is a frequency variable (wave number or spatial frequency).

[0063] Vertical height compensation: Raindrops exhibit fragmentation and merging effects, leading to differences in the noise spectrum between the ground and upper atmosphere (fragmentation (height > 140m) and merging (height < 140m) effects):

[0064] : Difference in height between the sensor and the cloud layer (m); Rainfall intensity is modeled using a hybrid Copula function. With noise power spectrum The nonlinear relationship is captured by the Gumbel function, which captures high-frequency tail correlations, while the Clayton function strengthens the correlation between low and medium probability events.

[0065] Marginal distribution function of rainfall intensity and noise spectrum; Weight parameters (optimized using Bayesian weighted average); (4) Dynamic suppression band Depth Spectral Subtraction Algorithm:

[0066] As a dynamic inhibitory factor:

[0067] To correct the parameters.

[0068] When the frequency band is in the characteristic band of blade cracks, zero suppression is implemented ( =0); when in the main energy region of rain noise. Sui Yuqiang Linear enhancement.

[0069] Feature extraction and matching are performed on the rain noise-suppressed acoustic fingerprint signal to achieve accurate fault identification. A preset fault acoustic fingerprint template library is loaded, which contains standard feature vectors of typical faults such as blade cracks, gearbox tooth breakage, and bearing inner ring wear. The dynamic time warping algorithm is used to calculate the matching degree between the feature vector to be detected and each fault feature vector in the template library. A matching degree threshold is set. When the matching degree is greater than or equal to the threshold, it is determined to be the corresponding fault type. The host computer triggers an audible and visual alarm and records the fault time, operating parameters, and original acoustic fingerprint data.

[0070] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, read-only optical discs, optical storage, etc.) containing computer-usable program code.

[0071] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0072] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0073] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0074] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the protection scope of the present invention.

Claims

1. A wind turbine acoustic signature adaptive detection method, characterized in that, Includes the following steps: Step S1: Use a sensor array to collect real-time acoustic signature signals from the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure. Step S2: Calculate temperature and humidity coupling correction parameters based on ambient humidity, and use ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula; Step S3: Construct a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm; Step S4: Using the depth spectrum subtraction algorithm, combined with the rain intensity noise spectrum mapping model and dynamic suppression factor, rain noise suppression is performed on the wind turbine acoustic signature signal to obtain a clean acoustic signature signal. Step S5: Match the cleaning acoustic signature signal with the preset wind turbine fault acoustic signature feature library to complete the wind turbine acoustic signature adaptive detection.

2. The wind turbine acoustic signature adaptive detection method according to claim 1, characterized in that, The sensor array mentioned in step S1 includes an acoustic fingerprint sensor, an optical rain sensor, and an environmental sensor. The acoustic fingerprint sensor is installed at the root of the wind turbine blades, the gearbox housing, and the middle section of the tower. The optical rain sensor is installed on the top of the wind turbine tower and the windward side of the blades of the wind turbine unit. The environmental sensors include a temperature and humidity sensor and a barometer; In step S1, the acoustic signature signal of the wind turbine is collected using an acoustic signature sensor, the raindrop diameter is collected by inversion using an optical rain sensor combined with infrared light shading effect, and the ambient humidity, ambient temperature and atmospheric pressure are monitored and collected using an environmental sensor.

3. The wind turbine acoustic signature adaptive detection method according to claim 2, characterized in that, The temperature and humidity coupling correction parameters calculated based on ambient humidity in step S2 are shown in the following formula: in, For temperature and humidity coupling correction parameters, For real-time ambient humidity, , , To correct the parameters; The air density correction formula in step S2 is shown below: in, For real-time air density, Relative humidity, Atmospheric pressure, For ambient temperature, This is the partial pressure of saturated water vapor.

4. The wind turbine acoustic signature adaptive detection method according to claim 3, characterized in that, Step S3, which involves constructing a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm, includes the following steps: Step S301: Combine rainfall intensity and average raindrop diameter to fit and determine the parameters of the raindrop diameter distribution function, and establish the raindrop diameter distribution function using the Gamma distribution; Step S302: Calculate the power spectrum of rain noise generated by raindrop impacting wind turbine components based on the raindrop kinetic energy distribution and real-time air density; Step S303: Based on the height difference between the sensor and the cloud layer, perform vertical height compensation on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights; Step S304: Use a hybrid Copula function to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum to obtain a rainfall intensity noise spectrum mapping model.

5. The wind turbine acoustic signature adaptive detection method according to claim 4, characterized in that, The hybrid Copula function described in step S304 includes at least a sub-Copula function for capturing high-frequency tail correlations and strengthening the association of low-probability events, and the function weights are optimized by Bayesian weighted average.

6. The wind turbine acoustic signature adaptive detection method according to claim 4, characterized in that, The raindrop diameter distribution function established in step S301 using the Gamma distribution is shown in the following equation: in, The diameter of the raindrop. For the intercept parameter, For shape parameters, The slope parameter; The calculation of the rain noise power spectrum generated by raindrop impacting wind turbine components based on the raindrop kinetic energy distribution and real-time air density in step S302 is shown in the following formula: in, The power spectrum of rain noise. The final velocity of the raindrop. The drag coefficient, For real-time air density, For the density of water, For correction factor, For frequency variables; In step S303, based on the height difference between the sensor and the cloud layer, vertical height compensation is performed on the rain noise power spectrum to correct the differences in rain noise power spectrum at different heights, as shown in the following formula: in, The difference in height between the sensor and the cloud layer; In step S304, a hybrid Copula function is used to model the nonlinear relationship between rainfall intensity and rainfall noise power spectrum, resulting in a rainfall intensity-noise spectrum mapping model, as shown in the following equation: in, Let be the edge distribution function of the power spectrum of rainfall intensity and rainfall noise. These are the weight parameters.

7. The wind turbine acoustic signature adaptive detection method according to claim 1, characterized in that, The dynamic suppression factor mentioned in step S4 combines the temperature and humidity coupling correction coefficient with the dynamic adjustment of rainfall intensity, and implements zero suppression in the fault characteristic frequency band of the wind turbine, while linearly increasing with rainfall intensity in the main energy area of ​​rain noise. The depth spectrum subtraction algorithm described in step S4 is shown in the following equation: in, As a dynamic inhibitor, To correct the parameters, Rain intensity.

8. A wind turbine acoustic signature adaptive detection system, characterized in that, include: The sensor acquisition module is used to acquire real-time data such as acoustic signature of the wind turbine, raindrop diameter, ambient humidity, ambient temperature, and atmospheric pressure using a sensor array. The correction parameter module is used to calculate temperature and humidity coupled correction parameters based on ambient humidity. It uses ambient humidity, ambient temperature and atmospheric pressure to correct the real-time air density according to the air density correction formula. A mapping model module is constructed to build a rain intensity noise spectrum mapping model based on raindrop diameter and real-time air density combined with a raindrop vertical breakup and merging effect compensation algorithm. The rain noise suppression module is used to suppress rain noise in the acoustic signature signal of the wind turbine by using a depth spectrum subtraction algorithm, combined with a rain intensity noise spectrum mapping model and a dynamic suppression factor, to obtain a clean acoustic signature signal. The adaptive acoustic signature detection module is used to match the clean acoustic signature signal with a preset acoustic signature feature library of wind turbine faults to complete the adaptive acoustic signature detection of wind turbines.

9. An electronic device, characterized in that, The method includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the wind turbine acoustic signature adaptive detection method according to any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the wind turbine acoustic signature adaptive detection method according to any one of claims 1-7.

Images