CN115982563A - Running state identification method, device and equipment for variable pitch bearing of wind turbine generator - Google Patents

Running state identification method, device and equipment for variable pitch bearing of wind turbine generator Download PDF

Info

Publication number
CN115982563A
CN115982563A CN202211501792.1A CN202211501792A CN115982563A CN 115982563 A CN115982563 A CN 115982563A CN 202211501792 A CN202211501792 A CN 202211501792A CN 115982563 A CN115982563 A CN 115982563A
Authority
CN
China
Prior art keywords
pitch bearing
wind turbine
audio
turbine generator
target
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202211501792.1A
Other languages
Chinese (zh)
Inventor
尚晓龙
童彤
张克功
郝延
李恭斌
杨立平
张小龙
陶成强
苏善斌
于满源
严兴成
魏昂昂
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huaneng Clean Energy Research Institute
Huaneng Jiuquan Wind Power Co Ltd
Original Assignee
Huaneng Clean Energy Research Institute
Huaneng Jiuquan Wind Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huaneng Clean Energy Research Institute, Huaneng Jiuquan Wind Power Co Ltd filed Critical Huaneng Clean Energy Research Institute
Priority to CN202211501792.1A priority Critical patent/CN115982563A/en
Publication of CN115982563A publication Critical patent/CN115982563A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02E—REDUCTION OF GREENHOUSE GAS [GHG] EMISSIONS, RELATED TO ENERGY GENERATION, TRANSMISSION OR DISTRIBUTION
    • Y02E10/00—Energy generation through renewable energy sources
    • Y02E10/70—Wind energy
    • Y02E10/72—Wind turbines with rotation axis in wind direction

Abstract

The disclosure provides a method, a device and equipment for identifying the running state of a variable pitch bearing of a wind turbine generator, wherein the method comprises the following steps: monitoring a variable-pitch bearing of the wind turbine generator by adopting a sound sensor to acquire an audio signal to be detected; performing feature extraction on the audio signal to be detected to obtain a first audio feature; and inputting the first audio features into the trained running state recognition model, and determining the target running state of the variable pitch bearing according to the output of the running state recognition model. Therefore, the running state of the variable pitch bearing can be automatically predicted according to the audio frequency characteristics of the audio frequency signal to be detected, which are obtained by monitoring the variable pitch bearing of the wind turbine generator, based on the deep learning technology.

Description

Running state identification method, device and equipment for variable pitch bearing of wind turbine generator
Technical Field
The disclosure relates to the technical field of artificial intelligence, in particular to a method, a device and equipment for identifying the running state of a variable pitch bearing of a wind turbine generator.
Background
The wind power variable pitch bearing is a key part of a variable pitch system in a wind turbine generator and is directly connected with blades of the wind turbine generator. In a severe operating environment of a wind turbine generator, the load borne by a wind turbine variable-pitch bearing is extremely complex, and the change of the operating state is easy to occur.
In order to obtain the operating state of the pitch bearing of the wind turbine generator in time and enable relevant workers to perform corresponding maintenance and repair on the pitch bearing of the wind turbine generator in time according to the operating state, so that normal operation of the wind turbine generator and safety of the workers are guaranteed, and how to identify the operating state of the pitch bearing of the wind turbine generator is very important.
Disclosure of Invention
The present disclosure provides a method, an apparatus, and a device for identifying an operating state of a pitch bearing of a wind turbine generator system, so as to solve at least one of technical problems in the related art to a certain extent. The technical scheme of the disclosure is as follows:
according to one aspect of the disclosure, a method for identifying an operation state of a pitch bearing of a wind turbine generator is provided, which includes:
monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor to obtain an audio signal to be detected;
performing feature extraction on the audio signal to be detected to obtain a first audio feature;
inputting the first audio characteristic into a trained operation state recognition model, and determining a target operation state of the variable pitch bearing according to the output of the operation state recognition model.
According to another aspect of the present disclosure, an operating state identification device for a pitch bearing of a wind turbine generator is provided, including:
the monitoring module is used for monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor so as to acquire an audio signal to be detected;
the extraction module is used for extracting the characteristics of the audio signal to be detected to obtain first audio characteristics;
and the determining module is used for inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model.
According to another aspect of the present disclosure, an electronic device is provided, which is characterized by comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, the method for identifying the operating state of the pitch bearing of the wind turbine set proposed by the above aspect of the present disclosure is implemented.
According to a further aspect of the present disclosure, a non-transitory computer-readable storage medium of computer instructions for causing a computer to perform the method for identifying an operational state of a pitch bearing of a wind turbine set forth in the above aspect of the present disclosure is provided.
According to a further aspect of the present disclosure, a computer program product is provided, which includes a computer program, and when the computer program is executed by a processor, the method for identifying an operating state of a pitch bearing of a wind turbine generator set proposed by the above aspect of the present disclosure is implemented.
The technical scheme provided by the embodiment of the disclosure at least brings the following beneficial effects:
monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor to obtain an audio signal to be detected; performing feature extraction on the audio signal to be detected to obtain a first audio feature; and inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model. Therefore, the running state of the variable pitch bearing can be automatically predicted according to the audio frequency characteristics of the audio frequency signal to be detected, which are obtained by monitoring the variable pitch bearing of the wind turbine generator, based on the deep learning technology.
Additional aspects and advantages of the disclosure will be set forth in part in the description which follows and, in part, will be obvious from the description, or may be learned by practice of the disclosure.
Drawings
The foregoing and/or additional aspects and advantages of the present disclosure will become apparent and readily appreciated from the following description of the embodiments, taken in conjunction with the accompanying drawings of which:
fig. 1 is a schematic flow chart of a method for identifying an operating state of a pitch bearing of a wind turbine provided in a first embodiment of the present disclosure;
fig. 2 is a schematic flow chart of a method for identifying an operating state of a pitch bearing of a wind turbine provided in a second embodiment of the present disclosure;
fig. 3 is a schematic flow chart of a method for identifying an operating state of a pitch bearing of a wind turbine provided in a third embodiment of the present disclosure;
fig. 4 is a schematic view illustrating an operation state identification process of a pitch bearing of a wind turbine provided by the present disclosure;
fig. 5 is a schematic structural diagram of an operation state identification device of a pitch bearing of a wind turbine provided in a fourth embodiment of the present disclosure;
FIG. 6 illustrates a block diagram of an exemplary electronic device suitable for use in implementing embodiments of the present invention.
Detailed Description
Reference will now be made in detail to the embodiments of the present disclosure, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the drawings are exemplary and intended to be illustrative of the present disclosure, and should not be construed as limiting the present disclosure.
In the related art, the method for identifying the running state of the variable-pitch bearing of the wind turbine generator is mainly based on a vibration signal. However, the running state of the pitch bearing of the wind turbine generator is variable, and since the vibration signal is difficult to transmit through the medium such as air, the sensor for acquiring the vibration signal is usually required to be deployed on the pitch bearing of the wind turbine generator, however, the deployment mode of the sensor is very easy to cause the transmission path of the vibration signal to be complicated, the deployment mode is complicated, the sensor is easy to fall off and fail after the long-time running of the pitch bearing of the wind turbine generator, and the running state identification of the pitch bearing of the wind turbine generator based on the vibration signal is complicated due to the problems.
In order to at least one problem, the present disclosure provides a method, an apparatus, and a device for identifying an operating state of a pitch bearing of a wind turbine generator.
The following describes a method, a device and equipment for identifying the operating state of a pitch bearing of a wind turbine generator according to an embodiment of the present disclosure with reference to the accompanying drawings.
Fig. 1 is a schematic flow diagram of a method for identifying an operating state of a pitch bearing of a wind turbine provided in an embodiment of the present disclosure.
The embodiment of the present disclosure is exemplified by the operating state identification method for a pitch bearing of a wind turbine generator being configured in an operating state identification device for a pitch bearing of a wind turbine generator, and the operating state identification device for a pitch bearing of a wind turbine generator can be applied to any electronic device, so that the electronic device can execute an operating state identification function for a pitch bearing of a wind turbine generator.
The electronic device may be any device having a computing capability, for example, a Personal Computer (PC), a mobile terminal, a server, and the like, and the mobile terminal may be a hardware device having various operating systems, touch screens, and/or display screens, such as a mobile phone, a tablet Computer, a Personal digital assistant, and a wearable device.
As shown in fig. 1, the method for identifying the operating state of the pitch bearing of the wind turbine generator may include the following steps:
step 101, monitoring a variable pitch bearing of a wind turbine generator by using a sound sensor to obtain an audio signal to be detected.
In this embodiment of the present disclosure, the audio signal to be detected may be an audio signal (which may also be referred to as a sound signal) obtained by monitoring a pitch bearing of the wind turbine generator by a sound sensor (such as a piezoelectric industrial sound pickup, a microphone, and the like) when the pitch bearing of the wind turbine generator operates.
It should be noted that there may be one or more sound sensors, and the present disclosure is not limited thereto.
It should also be noted that the deployment manner of the sound sensor is not limited in the present disclosure, for example, a plurality of sound sensors may be deployed in a plurality of different orientations of the pitch bearing of the wind turbine, or one sound sensor may be deployed in a certain orientation of the pitch bearing of the wind turbine, and the like.
As a possible implementation, a variety of different deployment schemes may be employed for the acoustic sensor; the noise reduction processing can be performed on the audio signals acquired in the different deployment schemes, and the signal-to-noise ratios of the audio signals acquired in the different deployment schemes after the noise reduction processing are respectively determined; according to the signal-to-noise ratio, the deployment scheme corresponding to the audio signal with the largest signal-to-noise ratio is determined from the audio signals acquired from the different deployment schemes, and the sound sensor can be deployed according to the deployment scheme.
It can be understood that the wind field environment is severe, the wind turbine generator set is internally provided with a complex electromagnetic environment, and the high-performance sound sensor can be comprehensively selected from the aspects of frequency response characteristics, sensitivity, magnetic shielding performance and the like of the sound sensor in order to ensure that the high-performance sound sensor has a good acquisition effect on signals of various frequency bands and ensure the quality of acquired audio signals to be detected.
In the embodiment of the disclosure, a pitch bearing of the wind turbine generator can be monitored by adopting the sound sensor so as to acquire an audio signal to be detected.
In a possible implementation manner of the embodiment of the disclosure, a sound sensor may be used to monitor a pitch bearing of a wind turbine generator to obtain an initial audio signal; the initial audio signal can be preprocessed to obtain an audio signal to be tested; wherein the pre-processing comprises at least one of noise reduction processing, continuity detection, endpoint detection, and signal enhancement.
In an embodiment of the present disclosure, the initial audio signal may be an audio signal obtained when a pitch bearing of the wind turbine is running.
In the embodiment of the present disclosure, the initial audio signal may be preprocessed to obtain the audio signal to be tested, where the preprocessing may include at least one of the following processing means:
1. noise reduction processing
In the embodiment of the present disclosure, the noise reduction processing may be used to remove noise or noise in the initial audio signal, for example, a method combining wiener filtering and wavelet packet analysis may be used to reduce noise in the initial audio signal, or a partial model recognition algorithm (for example, a VMD (variable Mode Decomposition) algorithm, a VME (variable Mode Extraction) algorithm, etc.) may be used to enhance the initial audio signal first, and then a method combining wiener filtering and wavelet packet analysis is used to reduce noise in the initial audio signal.
2. Continuity testing
In the disclosed embodiments, continuity detection may be used to detect whether the acquisition instants of the initial audio signal are continuous.
In the disclosed embodiments, the initial audio signal may include a plurality of sub audio signals. The sub-audio signal may be a segment of the original audio signal or a frame of the original audio signal.
It is understood that after the initial audio signal is obtained, a plurality of sub audio signals of the initial audio signal may be sequentially encapsulated and packed, and each data packet may have a corresponding time stamp. After the data packets corresponding to the multiple sub audio signals are obtained, whether the acquisition time corresponding to each sub audio signal in the initial audio signal is continuous or not can be determined according to the timestamps corresponding to the multiple data packets.
Therefore, in the method and the device, the audio signal to be detected can be determined according to the initial audio signal which is continuous at the acquisition time.
3. Endpoint detection
The endpoint Detection may be used to determine a start position and an end position of a desired signal from an initial audio signal, for example, the endpoint Detection of the initial audio signal may be implemented by using VAD (Voice Activity Detection) technology, a deep learning based Recurrent Neural Network (RNN) technology, and the like.
Therefore, in the present disclosure, the audio signal to be tested may be determined according to the audio signal between the start position and the end position in the initial audio signal.
4. Signal enhancement
In the embodiment of the present disclosure, the initial audio signal may be enhanced to extract a desired signal from the initial audio signal, for example, the initial audio signal may be enhanced by using the above-mentioned partial model identification algorithm (such as VMD algorithm, VME algorithm, etc.).
It should be noted that, in practical applications, the processing means or method for preprocessing the initial audio signal may be other, and the disclosure does not limit this.
Step 102, performing feature extraction on the audio signal to be detected to obtain a first audio feature.
In the embodiment of the present disclosure, feature extraction may be performed on an audio signal to be detected to obtain a first audio feature. For example, a principal component analysis algorithm may be used to perform feature extraction on the audio signal to be detected to obtain the first audio feature.
And 103, inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model.
In embodiments of the present disclosure, the target operational state may be indicative of an actual operational state of the pitch bearing.
The operation state may be, for example, a normal operation state, poor lubrication, structural damage, abnormal load, poor installation, and the like, which is not limited by the present disclosure.
Among them, lubrication failure, structural damage, abnormal load, and mounting failure, that is, relative to a normal operation state, may be collectively referred to as an abnormal operation state.
It should be noted that the target operating state may be one, or may also be multiple, and the disclosure does not limit this.
In an embodiment of the disclosure, the first audio feature may be input to a trained operating state recognition model to determine a target operating state of the pitch bearing from an output of the operating state recognition model.
In a possible implementation manner of the embodiment of the present disclosure, in response to a confirmation operation on a target operation state, the target operation state may be used to label the initial audio signal, or in response to a modification operation on the target operation state, the target operation state may be updated, and the updated target operation state may be used to label the initial audio signal, so as to obtain a sample audio signal; and may add the sample audio signal to the voiceprint database.
According to the method for identifying the running state of the variable-pitch bearing of the wind turbine generator, the variable-pitch bearing of the wind turbine generator is monitored by the sound sensor, so that an audio signal to be detected is obtained; performing feature extraction on the audio signal to be detected to obtain a first audio feature; and inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model. Therefore, the running state of the variable pitch bearing can be automatically predicted according to the audio frequency characteristics of the audio frequency signal to be detected, which are obtained by monitoring the variable pitch bearing of the wind turbine generator, based on the deep learning technology.
In order to clearly illustrate how the target operation state of the variable pitch bearing is determined according to the output of the operation state identification model in the above embodiments of the disclosure, the disclosure also provides an operation state identification method of the variable pitch bearing of the wind turbine generator.
Fig. 2 is a schematic flow chart of a method for identifying an operating state of a pitch bearing of a wind turbine provided in the second embodiment of the present disclosure.
As shown in fig. 2, the method for identifying the operating state of the pitch bearing of the wind turbine generator may include the following steps:
step 201, monitoring a pitch bearing of the wind turbine generator by using a sound sensor to acquire an audio signal to be detected.
Step 202, performing feature extraction on the audio signal to be detected to obtain a first audio feature.
The execution process of step 201 to step 202 may refer to the execution process of any embodiment of the present disclosure, and is not described herein again.
Step 203, inputting the first audio characteristic into the operation state identification model to obtain the prediction probabilities of the plurality of operation states output by the operation state identification model.
Wherein the predicted probability may be used to indicate the probability that the pitch bearing is in the corresponding operating state.
It should be noted that the explanation of the operation state in step 103 is also applicable to this embodiment, and is not described herein.
In the embodiment of the present disclosure, the first audio feature is input into the operation state identification model, and the prediction probabilities of the plurality of operation states output by the operation state identification model can be obtained.
And 204, determining a target operation state corresponding to the variable pitch bearing from the plurality of operation states according to the prediction probabilities of the plurality of operation states.
The explanation of the target operation state in step 103 is also applicable to this embodiment, and is not described herein again.
In the embodiment of the disclosure, according to the prediction probabilities of the multiple operation states, the target operation state corresponding to the pitch bearing can be determined from the multiple operation states.
As a possible implementation manner, for any one of the multiple operating states, a set probability threshold corresponding to any one of the operating states may be obtained, and the prediction probability of the any one of the operating states may be compared with the set probability threshold; when the prediction probability of any operation state is greater than the set probability threshold, the target operation state corresponding to the pitch bearing can be determined to comprise any operation state.
In the embodiment of the present disclosure, the set probability threshold may be preset, for example, may be 0.8, 0.9, and the like, which is not limited by the present disclosure.
In the embodiment of the present disclosure, any one of the plurality of operating states may have a corresponding set probability threshold, for example, when the operating state is a normal operating state, the corresponding set probability threshold may be 0.7; for another example, when the operation state is poor lubrication, the corresponding set probability threshold may be 0.9.
It should be noted that the above examples of the set probability threshold corresponding to the operating state are merely exemplary, and in practical applications, the set probability threshold corresponding to each operating state may be set as needed.
It should be further noted that the set probability threshold corresponding to each operating state may be the same or may be different, and the disclosure does not limit this.
As an example, the operation states include a normal operation state, a lubrication failure, a structural damage, an abnormal load and an installation failure, the set probability threshold corresponding to the normal operation state is 0.7, the prediction probability corresponding to the normal operation state is 0.2, the set probability threshold corresponding to the lubrication failure is 0.85, the prediction probability corresponding to the lubrication failure is 0.5, the set probability threshold corresponding to the structural damage is 0.7, the prediction probability corresponding to the structural damage is 0.9, the set probability threshold corresponding to the abnormal load is 0.85, the prediction probability corresponding to the abnormal load is 0.9, the set probability threshold corresponding to the installation failure is 0.85, and the prediction probability corresponding to the installation failure is 0.6.
As another possible implementation, a probability distribution may be generated according to the prediction probabilities of the plurality of operating states; the probability distribution is sampled (for example, randomly sampled), so that sampling probability can be obtained, and the operating state corresponding to the sampling probability can be determined as the target operating state corresponding to the pitch bearing.
As another possible implementation manner, the operating state corresponding to the maximum prediction probability among the prediction probabilities of the multiple operating states may be determined as the target operating state corresponding to the pitch bearing.
Therefore, the target operation state corresponding to the variable-pitch bearing of the wind turbine generator can be effectively determined based on the prediction probabilities of the multiple operation states. In addition, the target operation state corresponding to the variable pitch bearing of the wind turbine generator can be determined in different modes, and the flexibility and the applicability of the method can be improved.
In a possible implementation manner of the embodiment of the present disclosure, in response to a confirmation operation on the target running state, the first audio feature may be labeled by using the target running state, or in response to a modification operation on the target running state, the target running state may be updated, and the first audio feature may be labeled by using the updated target running state, so as to obtain the first training sample; and may add the first training sample to the voiceprint database.
Specifically, after the target operation state of the variable pitch bearing is determined, a user or related staff can check the operation state of the variable pitch bearing; under the condition that the actual running state of the variable-pitch bearing is matched with the target running state, the target running state can be confirmed; therefore, in the present disclosure, in response to a confirmation operation of a user or a related worker on a target operation state, the target operation state may be adopted to label the first audio feature to obtain a first training sample; or after the user or related staff checks the running state of the variable pitch bearing, under the condition that the difference exists between the actual running state and the target running state of the variable pitch bearing, the user or related staff can modify the target running state to update the target running state; therefore, in the present disclosure, the target running state may be updated in response to a modification operation of the user or a related worker on the target running state, and the first audio feature may be labeled by using the updated target running state to obtain the first training sample; further, the first training sample can be added to the voiceprint database to be used for training the running state recognition model subsequently, so that the prediction precision of the model is improved, namely the accuracy of the model prediction result is improved.
After the target operation state of the pitch bearing of the wind turbine is determined, in order to facilitate related workers to know the target operation state of the pitch bearing of the wind turbine in time, in yet another possible implementation manner of the embodiment of the disclosure, identification information of the wind turbine may be acquired, and first prompt information may be generated according to the identification information, where the first prompt information may be used to prompt the target operation state of the pitch bearing corresponding to the identification information, and may be displayed or sent.
In the embodiment of the present disclosure, the wind turbine may have corresponding identification information, where the identification information may be used to uniquely identify the corresponding wind turbine.
In another possible implementation manner of the embodiment of the present disclosure, positioning information of a pitch bearing of a wind turbine generator may be acquired; the positioning information can be used for prompting the position of the pitch bearing; and generating second prompt information according to the positioning information, wherein the second prompt information can be used for prompting the target running state of the pitch bearing of the wind turbine generator, the position of which is matched with the positioning information, and displaying or sending the second prompt information.
Therefore, related workers can conveniently acquire the running state of the variable pitch bearing of the wind turbine generator and the position of the variable pitch bearing in time so as to determine whether the variable pitch bearing of the wind turbine generator is abnormal or not, and the variable pitch bearing can be timely and pertinently overhauled and maintained under the condition that the abnormal variable pitch bearing of the wind turbine generator is determined.
According to the method for identifying the running state of the variable pitch bearing of the wind turbine generator, the first audio features are input into the running state identification model, so that the prediction probabilities of a plurality of running states output by the running state identification model are obtained; the prediction probability is used for indicating the probability that the variable-pitch bearing is in the corresponding operation state; and determining a target operation state corresponding to the variable pitch bearing from the plurality of operation states according to the prediction probabilities of the plurality of operation states. Therefore, the effective identification of the operation state of the variable pitch bearing of the wind turbine generator can be realized based on the prediction probabilities of the plurality of operation states output by the operation state identification model.
In order to clearly illustrate how the operation state identification model in any of the above embodiments of the present disclosure is obtained by training, the present disclosure also provides an operation state identification method for a pitch bearing of a wind turbine generator.
Fig. 3 is a schematic flow chart of a method for identifying an operating state of a pitch bearing of a wind turbine provided in the third embodiment of the present disclosure.
As shown in fig. 3, based on any of the above embodiments of the present disclosure, the method for identifying the operating state of the pitch bearing of the wind turbine generator may further include the following steps:
301, obtaining at least one target training sample from a voiceprint database; the target training sample is obtained by monitoring a first sample pitch bearing of a first sample wind turbine generator system by using a sound sensor.
It should be noted that the explanation of the sound sensor in step 101 is also applicable to this embodiment, and is not described herein again.
In the embodiment of the disclosure, the target training sample may be obtained by monitoring a pitch bearing of a sample wind turbine generator by using a sound sensor.
It should be noted that the target training sample may be an audio signal of the first sample pitch bearing of the first sample wind turbine generator system, which is acquired by the sound sensor when the first sample pitch bearing of the first sample wind turbine generator system is in a normal operation state; or the audio signal of the first sample pitch bearing of the first sample wind turbine generator set is acquired by the sound sensor when the first sample pitch bearing of the first sample wind turbine generator set is in an abnormal operation state, which is not limited by the disclosure.
In the embodiment of the present disclosure, the voiceprint database may be used to store training samples, so that in the present disclosure, a target training sample may be obtained from the voiceprint database, where the number of the target training samples may be one or multiple, and the present disclosure does not limit this.
In one possible implementation manner of the embodiment of the present disclosure, the voiceprint database may include at least one second training sample and at least one third training sample.
Wherein the second training sample may include a second audio feature labeled with a normal operating state; the second audio frequency characteristic is obtained by carrying out characteristic extraction on a first sample audio frequency signal, and the first sample audio frequency signal is obtained by monitoring a second sample variable pitch bearing in a second sample wind turbine generator set in a normal running state by adopting a sound sensor.
That is to say, a sound sensor is adopted to monitor a second sample pitch bearing in a second sample wind turbine generator set in a normal operation state, and a first sample audio signal can be obtained; performing feature extraction on the first sample audio signal to obtain a second audio feature; and marking the running state of the second audio characteristic to obtain a second training sample.
The third training sample can comprise a third audio feature and/or a fourth audio feature marked with an abnormal operation state; wherein the third audio feature may be obtained by feature extraction of a predicted audio signal, the predicted audio signal being generated based on the generation countermeasure network; the fourth audio characteristic can be obtained by performing characteristic extraction on a second sample audio signal, and the second sample audio signal can be obtained by monitoring a third sample pitch bearing in a third sample wind turbine generator set in an abnormal operation state by using a sound sensor.
In the embodiment of the present disclosure, the predicted audio signal may be generated based on a trained generated confrontation network (GAN), that is, the trained generator for generating confrontation network is used to generate the predicted audio signal, so as to implement data enhancement.
In a possible implementation manner of the embodiment of the disclosure, in order to obtain a trained generation countermeasure network, a noise feature, a fifth audio feature and a sixth audio feature may be obtained, where the fifth audio feature may be obtained by performing feature extraction on a third sample audio signal, and the third sample audio signal may be obtained by monitoring a fourth sample pitch bearing in a fourth sample wind turbine generator set in a normal operation state by using a sound sensor; the sixth audio characteristic can be obtained by extracting the characteristic of a fourth sample audio signal, and the fourth sample audio signal can be obtained by monitoring a fifth sample variable pitch bearing in a fifth sample wind turbine generator set in an abnormal operation state by using a sound sensor; fusing the noise characteristic and the fifth audio characteristic to obtain a fused characteristic; inputting the fusion characteristics into an initial generator for generating the countermeasure network to obtain seventh audio characteristics output by the generator; the seventh audio characteristic may be input to an initial discriminator generating the countermeasure network to obtain a first output value; the sixth audio feature may be input to an initial discriminator generating a counterpoise network to obtain a second output value; a first loss value may be determined based on the first output value and the second output value, such that the initial generative countermeasure network may be trained based on the first loss value to yield a trained generative countermeasure network.
In the embodiment of the present disclosure, a noise feature may be obtained, for example, a gaussian noise generating function may be used to generate the noise feature.
Note that the noise feature may have the same size as the fifth audio feature, for example, the fifth audio feature has a size of n × m, and the noise feature also has a size of n × m.
In the embodiment of the present disclosure, the noise feature and the fifth audio feature may be fused to obtain a fusion feature, for example, the noise feature and the fifth audio feature may be added to obtain the fusion feature; for another example, the noise feature and the fifth audio feature may be spliced to obtain a fusion feature.
In an embodiment of the disclosure, the seventh audio characteristic may be input to the discriminator of the initial generation countermeasure network to obtain the first output value, for example, the first output value is D (X), where X is the seventh audio characteristic.
In the disclosed embodiment, the sixth audio feature is input to the initial discriminator of the generative confrontation network to obtain the second output value, for example, the second output value is D (Y), where Y is the sixth audio feature.
In the embodiment of the present disclosure, based on the first output value and the second output value, the first loss value may be determined, for example, the first output value is D (X), the second output value is D (Y), and the first loss value V may be determined according to the following formula:
Figure BDA0003968007210000141
wherein X is a seventh audio feature, Y is a sixth audio feature, D is an initial challenge network generating evaluator, G is an initial challenge network generating evaluator, and E (-) is a distribution function expectation.
In the disclosed embodiment, the initial generative countermeasure network may be trained according to the first loss value to obtain a trained generative countermeasure network. For example, based on the first loss value, model parameters of the generator and the discriminator that initially generate the countermeasure network may be adjusted to minimize a value of the first loss value.
It should be noted that, the above is only exemplified by taking the termination condition of the training for generating the countermeasure network as the minimization of the first loss value, and in practical application, other termination conditions may also be set, for example, the number of times of training reaches the set number, the training duration reaches the set duration, the loss value converges, and the like, which is not limited by the present disclosure.
Therefore, training of the generation of the countermeasure network can be realized, and the accuracy and reliability of the generation result of the countermeasure network can be improved.
It should be noted that, by using the predicted audio signal generated by the trained generator for generating the countermeasure network, the operating state corresponding to the predicted audio signal may be an abnormal operating state of a fifth sample pitch bearing in a fifth sample wind turbine generator corresponding to a fourth sample audio signal used during initial generation of the countermeasure network.
Step 302, performing running state prediction on the target training sample based on the initial running state recognition model to obtain output probabilities of a plurality of running states output by the initial running state recognition model.
It should be noted that the explanation of the operation state in step 103 is also applicable to this embodiment, and is not described herein again.
In the embodiment of the disclosure, the initial operation state recognition model is adopted to predict the operation state of the target training sample, so that the output probabilities of a plurality of operation states output by the initial operation state recognition model can be obtained.
Step 303, determining a predicted operating state from the plurality of operating states according to the output probabilities of the plurality of operating states.
In the disclosed embodiment, the predicted operating state may be determined from a plurality of operating states based on output probabilities of the plurality of operating states.
It should be noted that the method for determining the target operation state corresponding to the pitch bearing in step 204 is similar to the method for determining the predicted operation state in this embodiment, and details are not repeated here.
And step 304, training the initial operation state recognition model according to the difference between the predicted operation state and the marked operation state marked by the target training sample to obtain a trained operation state recognition model.
In the embodiment of the disclosure, the marked operation state may be used to indicate an actual operation state corresponding to a pitch bearing of the sample wind turbine generator when the target training sample is collected.
In the embodiment of the disclosure, the initial operation state recognition model may be trained according to a difference between the predicted operation state and the labeled operation state labeled by the target training sample, so that the trained operation state recognition model may be obtained.
As an example, a second loss value may be generated according to a difference between the predicted operating state and an labeled operating state labeled by the target training sample, and a model parameter in the initial fault identification model may be adjusted according to the second loss value to minimize the second loss value, where the difference between the predicted operating state and the labeled operating state labeled by the target training sample is in a positive correlation with the second loss value, that is, the smaller the difference is, the smaller the second loss value is, and conversely, the larger the difference is, the larger the second loss value is.
It should be noted that, the above example is performed by taking only the termination condition of the model training as the minimization of the loss value, and in practical application, other termination conditions may also be set, for example, the number of times of training reaches the set number of times, the training duration reaches the set duration, the loss value converges, and the like, which is not limited by the present disclosure.
According to the method for identifying the running state of the variable pitch bearing of the wind turbine generator, at least one target training sample is obtained from a voiceprint database; the target training sample is obtained by monitoring a variable pitch bearing of a sample wind turbine generator by adopting a sound sensor; predicting the running state of the target training sample based on the initial running state recognition model to obtain the output probability of a plurality of running states output by the initial running state recognition model; determining a predicted operating state from the plurality of operating states according to the output probabilities of the plurality of operating states; and training the initial running state recognition model according to the difference between the predicted running state and the marked running state marked by the target training sample to obtain the trained running state recognition model. Therefore, the trained operation state recognition model is obtained by training the initial operation state recognition model in advance, so that the operation state of the variable pitch bearing of the wind turbine generator can be predicted by adopting the trained operation state recognition model, and the accuracy of a prediction result can be improved.
As an application scenario, the method for identifying the operating state of the pitch bearing of the wind turbine generator is applied to an operating state identification device of the pitch bearing of the wind turbine generator for explanation, and an operating state identification process of the pitch bearing of the wind turbine generator is shown in fig. 4.
S1: and establishing a voiceprint database of a variable pitch bearing of the wind turbine generator.
In the method, sample voiceprint voice data (recorded as sample audio signals in the method) of a pitch bearing of the wind turbine generator in a normal operation state and in an abnormal operation state (namely poor lubrication, structural damage, abnormal load, poor installation and the like) can be collected as much as possible, the sample voiceprint voice data cover the number of operation state types, and the data volume can effectively guarantee the type of the operation state which can be identified by the operation state identification model and the accuracy of the identification result.
It can be understood that the voiceprint voice data of the wind turbine generator system with the variable-pitch bearing in the abnormal operation state are less, and the voiceprint voice data of the wind turbine generator system with the variable-pitch bearing in the abnormal operation state can be automatically generated through a generation countermeasure network algorithm, so that the abnormal data in the voiceprint voice database can be enriched through the algorithm.
S2: and establishing an operation state identification model based on the voiceprint database.
The operation state identification model can establish a mapping relation between a voiceprint feature set of a variable pitch bearing of the wind turbine generator and an operation state based on an artificial intelligent algorithm such as deep learning.
S3: and deploying the operation state identification model to an operation state identification device of a variable pitch bearing of the wind turbine generator on a monitoring site.
The running state identification device of the variable pitch bearing of the wind turbine generator set can comprise a monitoring module, an extraction module and a determination module:
the monitoring module is used for monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor (such as a piezoelectric industrial sound pickup) so as to acquire an initial audio signal; preprocessing the initial audio signal to obtain an audio signal to be detected; wherein the pre-processing comprises at least one of noise reduction processing, continuity detection, endpoint detection, and signal enhancement.
The extraction module is configured to perform feature extraction on the audio signal to be detected to obtain a first audio feature, for example, a principal component analysis algorithm may be adopted to perform feature extraction on the audio signal to be detected to obtain the first audio feature.
And the determining module is used for inputting the first audio features into the trained running state recognition model so as to determine the target running state of the variable pitch bearing according to the output of the running state recognition model, namely establishing the mapping relation between the first audio features and the running state of the variable pitch bearing of the wind turbine generator.
It should be noted that, in order to ensure the quality of the acquired initial audio signal or the audio signal to be detected, a high-performance sound sensor can be comprehensively selected from the aspects of frequency response characteristics, sensitivity, magnetic shielding performance and the like, and the sound sensor can bear a relatively complex electromagnetic environment inside the wind turbine generator and has a good acquisition effect on signals of various frequency bands, so as to accurately acquire the initial audio signal when the pitch bearing operates.
S4: and deploying the running state recognition device of the variable pitch bearing of the wind turbine generator at the corresponding position around the variable pitch bearing to be detected of the wind turbine generator to be detected.
The running state recognition device of the variable pitch bearing of at least one set of wind turbine generator set is deployed on one variable pitch bearing to be tested, the running state recognition devices of the variable pitch bearings of all the wind turbine generator sets can be deployed on the periphery of the variable pitch bearing to be tested, and are not in any electrical connection and physical contact with the variable pitch bearing to be tested, so that the running state of the variable pitch bearing to be tested is prevented from being interfered.
The method can be used for comparing the influence of different deployment schemes on the operation state identification result, and determining the operation state identification device deployment scheme of the variable pitch bearing of the wind turbine generator set, which has strong anti-interference performance and high identification precision.
S5: the running state recognition device of the variable-pitch bearing of the wind turbine generator set can output a running state recognition result and feed back the running state recognition result to a control center of a wind power plant station or a regional centralized control center of a higher level.
When the operation state identification result is sent to the wind power plant station, the operation state identification result can be sent to the regional relevant centralized control center through the special network.
S6: operation and maintenance personnel of the wind power plant can confirm the field state, can supplement the newly acquired initial audio signal and the corresponding running state into a voiceprint database, and can optimize the running state recognition model.
It should be noted that the optimized operation state identification model may perform update iteration on the initial operation state identification model in an online or offline manner.
According to the method for identifying the running state of the variable-pitch bearing of the wind turbine generator, the running state of the variable-pitch bearing of the wind turbine generator can be identified based on the sound signal, obtained by the sound sensor, of the variable-pitch bearing of the wind turbine generator during running; in the method for identifying the running state of the variable pitch bearing of the wind turbine generator, the arrangement mode of the sound sensor is simple, direct contact with the variable pitch bearing to be detected of the wind turbine generator to be detected is not needed, and the sound sensor can be flexibly arranged at the peripheral position of the variable pitch bearing to be detected of the wind turbine generator to be detected.
Corresponding to the method for identifying the operating state of the pitch bearing of the wind turbine provided in the embodiments of fig. 1 to 3, the present disclosure also provides a device for identifying the operating state of the pitch bearing of the wind turbine, and since the device for identifying the operating state of the pitch bearing of the wind turbine provided in the embodiments of the present disclosure corresponds to the method for identifying the operating state of the pitch bearing of the wind turbine provided in the embodiments of fig. 1 to 3, the embodiment of the method for identifying the operating state of the pitch bearing of the wind turbine is also applicable to the device for identifying the operating state of the pitch bearing of the wind turbine provided in the embodiments of the present disclosure, and will not be described in detail in the embodiments of the present disclosure.
Fig. 5 is a schematic structural diagram of an operation state identification device of a pitch bearing of a wind turbine provided in the fourth embodiment of the present disclosure.
As shown in fig. 5, the device 500 for identifying the operating state of the pitch bearing of the wind turbine may include: a monitoring module 501, an extraction module 502, and a determination module 503.
The monitoring module 501 is configured to monitor a pitch bearing of the wind turbine generator by using a sound sensor to obtain an audio signal to be detected.
The extracting module 502 is configured to perform feature extraction on the audio signal to be detected to obtain a first audio feature.
A determining module 503, configured to input the first audio feature into the trained operation state recognition model, so as to determine a target operation state of the pitch bearing according to an output of the operation state recognition model.
In a possible implementation manner of the embodiment of the present disclosure, the determining module 503 is configured to: inputting the first audio features into the operation state recognition model to obtain the prediction probabilities of a plurality of operation states output by the operation state recognition model; wherein the predicted probability is used to indicate the probability that the pitch bearing is in the corresponding operating state; and determining a target operation state corresponding to the variable pitch bearing from the plurality of operation states according to the prediction probabilities of the plurality of operation states.
In a possible implementation manner of the embodiment of the present disclosure, the determining module 503 is configured to: aiming at any one of the multiple running states, acquiring a set probability threshold corresponding to the running state; comparing the prediction probability of any operation state with a set probability threshold; in response to the fact that the prediction probability of any operation state is larger than a set probability threshold, determining that the target operation state corresponding to the variable-pitch bearing comprises any operation state; or generating probability distribution according to the prediction probability of a plurality of operation states; sampling the probability distribution to obtain a sampling probability; determining the running state corresponding to the sampling probability as a target running state corresponding to the variable-pitch bearing; or determining the operation state corresponding to the maximum prediction probability in the prediction probabilities of the multiple operation states as the target operation state corresponding to the variable-pitch bearing.
In a possible implementation manner of the embodiment of the present disclosure, the monitoring module 501 is configured to: monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor to obtain an initial audio signal; preprocessing the initial audio signal to obtain an audio signal to be detected; wherein the pre-processing comprises at least one of noise reduction processing, continuity detection, endpoint detection, and signal enhancement.
In a possible implementation manner of the embodiment of the present disclosure, the device 500 for identifying an operating state of a pitch bearing of a wind turbine generator may further include:
and the processing module is used for responding to the operation of confirming the target running state, adopting the target running state to label the first audio characteristic, or responding to the operation of modifying the target running state, updating the target running state, and adopting the updated target running state to label the first audio characteristic so as to obtain the first training sample.
And the adding module is used for adding the first training sample into the voiceprint database.
In a possible implementation manner of the embodiment of the present disclosure, the determining module 503 is configured to: acquiring identification information of a wind turbine of the wind turbine; generating first prompt information according to the identification information, wherein the first prompt information is used for prompting the running state of the variable pitch bearing corresponding to the identification information; displaying or sending first prompt information; or acquiring positioning information of a variable-pitch bearing of the wind turbine generator; the positioning information is used for prompting the position of the variable-pitch bearing; generating second prompt information according to the positioning information, wherein the second prompt information is used for prompting the running state of a variable pitch bearing of the wind turbine generator, and the position of the second prompt information is matched with the positioning information; and displaying or sending the second prompt message.
In a possible implementation manner of the embodiment of the present disclosure, the operation state recognition model is obtained by training through the following steps: obtaining at least one target training sample from a voiceprint database; the target training sample is obtained by monitoring a first sample variable-pitch bearing of a first sample wind turbine generator system by using a sound sensor; predicting the running state of the target training sample based on the initial running state recognition model to obtain the output probability of a plurality of running states output by the initial running state recognition model; determining a predicted operating state from the plurality of operating states according to the output probabilities of the plurality of operating states; and training the initial running state recognition model according to the difference between the predicted running state and the marked running state marked by the target training sample to obtain the trained running state recognition model.
In a possible implementation manner of the embodiment of the present disclosure, the voiceprint database includes at least one second training sample and at least one third training sample; the second training sample comprises a second audio characteristic marked with a normal running state; the second audio frequency characteristic is obtained by extracting the characteristic of a first sample audio frequency signal, and the first sample audio frequency signal is obtained by monitoring a second sample variable pitch bearing in a second sample wind turbine generator set in a normal operation state by adopting a sound sensor; the third training sample comprises a third audio feature and/or a fourth audio feature marked with an abnormal operating state; the third audio feature is obtained by feature extraction of a predicted audio signal, and the predicted audio signal is generated based on a trained generation countermeasure network; the fourth audio characteristic is obtained by extracting the characteristic of a second sample audio signal, and the second sample audio signal is obtained by monitoring a third sample variable pitch bearing in a third sample wind turbine generator set in an abnormal operation state by using a sound sensor.
According to the operation state recognition device of the variable pitch bearing of the wind turbine generator, the variable pitch bearing of the wind turbine generator is monitored by the sound sensor, so that an audio signal to be detected is obtained; performing feature extraction on the audio signal to be detected to obtain a first audio feature; and inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model. Therefore, the running state of the variable pitch bearing can be automatically predicted according to the audio frequency characteristics of the audio frequency signal to be detected, which is obtained by monitoring the variable pitch bearing of the wind turbine generator system, based on the deep learning technology.
In order to implement the foregoing embodiments, the present disclosure further provides an electronic device, which is characterized by including a memory, a processor, and a computer program that is stored in the memory and is executable on the processor, and when the processor executes the program, the method for identifying the operating state of the pitch bearing of the wind turbine generator set according to any one of the foregoing embodiments of the present disclosure is implemented.
In order to achieve the above embodiments, the present disclosure further proposes a non-transitory computer readable storage medium, on which a computer program is stored, and when the program is executed by a processor, the method for identifying an operating state of a pitch bearing of a wind turbine generator set is implemented as proposed in any one of the foregoing embodiments of the present disclosure.
In order to achieve the above embodiments, the present disclosure further provides a computer program product, and when instructions in the computer program product are executed by a processor, the method for identifying an operating state of a pitch bearing of a wind turbine generator set according to any one of the foregoing embodiments of the present disclosure is performed.
As shown in fig. 6, electronic device 12 is in the form of a general purpose computing device. The components of the electronic device 12 may include, but are not limited to: one or more processors or processing units 16, a system memory 28, and a bus 18 that couples various system components including the system memory 28 and the processing unit 16.
Bus 18 represents one or more of any of several types of bus structures, including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, and a processor or local bus using any of a variety of bus architectures. These architectures include, but are not limited to, industry Standard Architecture (ISA) bus, micro Channel Architecture (MAC) bus, enhanced ISA bus, video Electronics Standards Association (VESA) local bus, and Peripheral Component Interconnect (PCI) bus, to name a few.
Electronic device 12 typically includes a variety of computer system readable media. Such media may be any available media that is accessible by electronic device 12 and includes both volatile and nonvolatile media, removable and non-removable media.
Memory 28 may include computer system readable media in the form of volatile Memory, such as Random Access Memory (RAM) 30 and/or cache Memory 32. The electronic device 12 may further include other removable/non-removable, volatile/nonvolatile computer system storage media. By way of example only, storage system 34 may be used to read from and write to non-removable, nonvolatile magnetic media (not shown in FIG. 6, commonly referred to as a "hard drive"). Although not shown in FIG. 6, a disk drive for reading from and writing to a removable, nonvolatile magnetic disk (e.g., a "floppy disk") and an optical disk drive for reading from or writing to a removable, nonvolatile optical disk (e.g., a Compact disk Read Only Memory (CD-ROM), a Digital versatile disk Read Only Memory (DVD-ROM), or other optical media) may be provided. In these cases, each drive may be connected to bus 18 by one or more data media interfaces. Memory 28 may include at least one program product having a set (e.g., at least one) of program modules that are configured to carry out the functions of embodiments of the disclosure.
A program/utility 40 having a set (at least one) of program modules 42 may be stored, for example, in memory 28, such program modules 42 including but not limited to an operating system, one or more application programs, other program modules, and program data, each of which or some combination of which may comprise an implementation of a network environment. Program modules 42 generally perform the functions and/or methodologies of the embodiments described in this disclosure.
Electronic device 12 may also communicate with one or more external devices 14 (e.g., keyboard, pointing device, display 24, etc.), with one or more devices that enable a user to interact with electronic device 12, and/or with any devices (e.g., network card, modem, etc.) that enable electronic device 12 to communicate with one or more other computing devices. Such communication may be through an input/output (I/O) interface 22. Also, the electronic device 12 may communicate with one or more networks (e.g., a Local Area Network (LAN), a Wide Area Network (WAN), and/or a public Network such as the Internet) via the Network adapter 20. As shown, the network adapter 20 communicates with other modules of the electronic device 12 via the bus 18. It should be understood that although not shown in the figures, other hardware and/or software modules may be used in conjunction with electronic device 12, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems, among others.
The processing unit 16 executes various functional applications and data processing, for example, implementing the methods mentioned in the foregoing embodiments, by running a program stored in the system memory 28.
In the description herein, references to the description of the term "one embodiment," "some embodiments," "an example," "a specific example," or "some examples," etc., mean that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present disclosure. In this specification, the schematic representations of the terms used above are not necessarily intended to refer to the same embodiment or example. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples. Furthermore, various embodiments or examples and features of different embodiments or examples described in this specification can be combined and combined by one skilled in the art without contradiction.
Furthermore, the terms "first", "second" and "first" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to implicitly indicate the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of the present disclosure, "plurality" means at least two, e.g., two, three, etc., unless explicitly defined otherwise.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing steps of a custom logic function or process, and alternate implementations are included within the scope of the preferred embodiment of the present disclosure in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the embodiments of the present disclosure.
The logic and/or steps represented in the flowcharts or otherwise described herein, e.g., an ordered listing of executable instructions that can be considered to implement logical functions, can be embodied in any computer-readable medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions. For the purposes of this description, a "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer-readable medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, the computer-readable medium could even be paper or another suitable medium upon which the program is printed, as the program can be electronically captured, via for instance optical scanning of the paper or other medium, then compiled, interpreted or otherwise processed in a suitable manner if necessary, and then stored in a computer memory.
It should be understood that portions of the present disclosure may be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, any one or combination of the following techniques, which are well known in the art, may be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application specific integrated circuit having an appropriate combinational logic gate circuit, a Programmable Gate Array (PGA), a Field Programmable Gate Array (FPGA), or the like.
It will be understood by those skilled in the art that all or part of the steps carried by the method for implementing the above embodiments may be implemented by hardware related to instructions of a program, which may be stored in a computer readable storage medium, and when the program is executed, the program includes one or a combination of the steps of the method embodiments.
In addition, functional units in the embodiments of the present disclosure may be integrated into one processing module, or each unit may exist alone physically, or two or more units are integrated into one module. The integrated module can be realized in a hardware mode, and can also be realized in a software functional module mode. The integrated module, if implemented in the form of a software functional module and sold or used as a stand-alone product, may also be stored in a computer readable storage medium.
The storage medium mentioned above may be a read-only memory, a magnetic or optical disk, etc. While embodiments of the present disclosure have been shown and described above, it will be understood that the above embodiments are exemplary and not to be construed as limiting the present disclosure, and that changes, modifications, substitutions and alterations may be made to the above embodiments by those of ordinary skill in the art within the scope of the present disclosure.

Claims (10)

1. A method for identifying the running state of a variable-pitch bearing of a wind turbine generator is characterized by comprising the following steps:
monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor to obtain an audio signal to be detected;
extracting the characteristics of the audio signal to be detected to obtain first audio characteristics;
inputting the first audio features into a trained operation state recognition model, and determining a target operation state of the variable pitch bearing according to the output of the operation state recognition model.
2. The method of claim 1, wherein inputting the first audio feature into a trained operating state recognition model to determine a target operating state of the pitch bearing from an output of the operating state recognition model comprises:
inputting the first audio characteristic into the operation state identification model to obtain the prediction probability of a plurality of operation states output by the operation state identification model; wherein the predicted probability is used to indicate a probability that the pitch bearing is in a corresponding operating state;
and determining a target operation state corresponding to the variable-pitch bearing from the plurality of operation states according to the prediction probabilities of the plurality of operation states.
3. The method of claim 2, wherein said determining a target operating state from the plurality of operating states corresponding to the pitch bearing based on the predicted probabilities of the plurality of operating states comprises:
aiming at any one of the multiple operation states, acquiring a set probability threshold corresponding to the any operation state;
comparing the predicted probability of any one of the operating states with the set probability threshold;
in response to the prediction probability of any operation state being larger than the set probability threshold, determining that the target operation state corresponding to the variable-pitch bearing comprises any operation state;
or the like, or, alternatively,
generating probability distribution according to the prediction probabilities of the plurality of operation states;
sampling the probability distribution to obtain a sampling probability;
determining the running state corresponding to the sampling probability as a target running state corresponding to the variable-pitch bearing;
or the like, or a combination thereof,
and determining the operation state corresponding to the maximum prediction probability in the prediction probabilities of the plurality of operation states as the target operation state corresponding to the variable-pitch bearing.
4. The method according to claim 1, wherein the monitoring of the pitch bearing of the wind turbine generator with the sound sensor to obtain the audio signal to be measured comprises:
monitoring a variable pitch bearing of the wind turbine generator by adopting the sound sensor to obtain an initial audio signal;
preprocessing the initial audio signal to obtain the audio signal to be detected;
wherein the pre-processing comprises at least one of noise reduction processing, continuity detection, endpoint detection, and signal enhancement.
5. The method of claim 1, further comprising:
in response to the operation of confirming the target running state, labeling the first audio feature by using the target running state, or in response to the operation of modifying the target running state, updating the target running state, and labeling the first audio feature by using the updated target running state to obtain a first training sample;
adding the first training sample to a voiceprint database.
6. The method of claim 1, wherein after inputting the first audio feature into a trained operating state recognition model to determine a target operating state of the pitch bearing from an output of the operating state recognition model, the method further comprises:
acquiring identification information of a wind turbine of the wind turbine;
generating first prompt information according to the identification information, wherein the first prompt information is used for prompting the running state of the variable-pitch bearing corresponding to the identification information;
displaying or sending the first prompt message;
alternatively, the first and second electrodes may be,
acquiring positioning information of a variable-pitch bearing of the wind turbine generator; the positioning information is used for prompting the position of the variable pitch bearing;
generating second prompt information according to the positioning information, wherein the second prompt information is used for prompting the running state of a variable pitch bearing of the wind turbine generator, the position of which is matched with the positioning information;
and displaying or sending the second prompt message.
7. The method according to any one of claims 1 to 6, wherein the operating state recognition model is trained using the steps of:
obtaining at least one target training sample from a voiceprint database; the target training sample is obtained by monitoring a first sample variable-pitch bearing of a first sample wind turbine generator system by using a sound sensor;
predicting the running state of the target training sample based on an initial running state recognition model to obtain the output probability of a plurality of running states output by the initial running state recognition model;
determining a predicted operating state from the plurality of operating states based on the output probabilities of the plurality of operating states;
and training the initial running state recognition model according to the difference between the predicted running state and the marked running state marked by the target training sample to obtain a trained running state recognition model.
8. The method of claim 7, wherein the voiceprint database includes at least one second training sample and at least one third training sample;
the second training sample comprises a second audio feature labeled with a normal operating state; the second audio frequency characteristic is obtained by extracting the characteristic of a first sample audio frequency signal, and the first sample audio frequency signal is obtained by monitoring a second sample variable pitch bearing in a second sample wind turbine generator set in a normal operation state by adopting a sound sensor;
the third training sample comprises a third audio feature and/or a fourth audio feature labeled with an abnormal operating state; wherein the third audio feature is obtained by feature extraction of a predicted audio signal, and the predicted audio signal is generated based on a trained generative confrontation network;
the fourth audio characteristic is obtained by extracting the characteristic of a second sample audio signal, and the second sample audio signal is obtained by monitoring a third sample variable pitch bearing in a third sample wind turbine generator set in an abnormal operation state by adopting a sound sensor.
9. An operating state recognition device of a variable pitch bearing of a wind turbine generator, characterized in that the device comprises:
the monitoring module is used for monitoring a variable pitch bearing of the wind turbine generator by adopting a sound sensor so as to acquire an audio signal to be detected;
the extraction module is used for extracting the characteristics of the audio signal to be detected to obtain first audio characteristics;
and the determining module is used for inputting the first audio characteristic into the trained operation state recognition model so as to determine the target operation state of the variable pitch bearing according to the output of the operation state recognition model.
10. An electronic device, comprising a memory, a processor and a computer program stored on the memory and executable on the processor, wherein the processor, when executing the program, implements the method for identifying an operating state of a pitch bearing of a wind turbine according to any one of claims 1 to 8.
CN202211501792.1A 2022-11-28 2022-11-28 Running state identification method, device and equipment for variable pitch bearing of wind turbine generator Pending CN115982563A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202211501792.1A CN115982563A (en) 2022-11-28 2022-11-28 Running state identification method, device and equipment for variable pitch bearing of wind turbine generator

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202211501792.1A CN115982563A (en) 2022-11-28 2022-11-28 Running state identification method, device and equipment for variable pitch bearing of wind turbine generator

Publications (1)

Publication Number Publication Date
CN115982563A true CN115982563A (en) 2023-04-18

Family

ID=85968885

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202211501792.1A Pending CN115982563A (en) 2022-11-28 2022-11-28 Running state identification method, device and equipment for variable pitch bearing of wind turbine generator

Country Status (1)

Country Link
CN (1) CN115982563A (en)

Similar Documents

Publication Publication Date Title
Scanlon et al. Residual life prediction of rotating machines using acoustic noise signals
Soualhi et al. Hidden Markov models for the prediction of impending faults
CN111459700B (en) Equipment fault diagnosis method, diagnosis device, diagnosis equipment and storage medium
Soualhi et al. Prognosis of bearing failures using hidden Markov models and the adaptive neuro-fuzzy inference system
Janjarasjitt et al. Bearing condition diagnosis and prognosis using applied nonlinear dynamical analysis of machine vibration signal
JP6374466B2 (en) Sensor interface device, measurement information communication system, measurement information communication method, and measurement information communication program
CA2874991A1 (en) Methods and apparatuses for defect diagnosis in a mechanical system
Krause et al. Damage detection for wind turbine rotor blades using airborne sound
WO2021027213A1 (en) Detection method and apparatus, electronic device and computer-readable medium
CN108304348B (en) Bearing residual life prediction method based on binary wiener process
CN113670434B (en) Method and device for identifying sound abnormality of substation equipment and computer equipment
US11747405B2 (en) System and method for audio and vibration based power distribution equipment condition monitoring
JP2019067197A (en) Method for detecting trouble sign
EP3759558A1 (en) Intelligent audio analytic apparatus (iaaa) and method for space system
KR20210006832A (en) Method and apparatus for machine fault diagnosis
JP2018189522A (en) Method for diagnosing sign of facility failure
CN116415126A (en) Method, device and computing equipment for anomaly detection of doctor blades of paper machine
CN108333443B (en) Method for alarming intermittent defects of power equipment
CN114242085A (en) Fault diagnosis method and device for rotating equipment
CN115982563A (en) Running state identification method, device and equipment for variable pitch bearing of wind turbine generator
CN116907029A (en) Method for detecting abnormality of fan in outdoor unit, control device and air conditioner outdoor unit
CN113805105A (en) Three-phase transformer detection method and system
EP4254261A1 (en) Blade fault diagnosis method, apparatus and system, and storage medium
CN114626400A (en) Equipment fault diagnosis device, equipment fault diagnosis method, intelligent factory system and application agent
CN114255373B (en) Sequence anomaly detection method, device, electronic equipment and readable medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination