Wear state determination method and device, storage medium and electronic device

Abstract

The invention discloses a wear state determination method and device, a storage medium and an electronic device.The method comprises the steps that first wear state data of a first brake pad are obtained, and the first wear state data at least comprise a vibration signal and a temperature signal corresponding to the first brake pad; determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad; and the first wear state feature vector is input into a trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector. By means of the method, the problem that in the related technology, due to uneven abrasion of the brake pad, operation of the wind generating set is unstable can be solved.

Description

Technical Field

[0001] The present application relates to the field of wind power, and in particular, to a method and device for determining a wear state, a storage medium, and an electronic device. Background Art

[0002] Frequent operation of the yaw brake of a wind turbine can easily lead to wear of the brake pads, which in turn affects the safe and stable operation of the wind turbine. This problem is particularly prominent for wind turbines installed in coastal areas where typhoons are frequent.

[0003] The hydraulic active disc yaw brake of a wind turbine usually consists of two halves of clamps, four piston heads, two brake pads and a hydraulic drive system. Through hydraulic drive, the piston heads on the upper and lower clamps and the brake pads connected to them are pushed to tightly clamp the brake disc, thereby achieving the purpose of yaw braking. After the unit has been operating normally for half a year, the yaw brake disc of the wind turbine will have long grooves, small pits, convex spots, etc., there will be several long grooves on the edge of the yaw brake disc, there will be powdery flakes of metal debris on the upper surface of the yaw brake disc, and the hydraulic station will have a low oil level fault alarm. After replacing one of the brake pads and rechecking it, it will be found that the new brake pads will soon be severely worn out.

[0004] Due to the small installation space of the yaw brake and the limited external dimensions of the brake, the brake cylinder body is impacted by the centrifugal force of the unit rotation, which causes the piston stroke inside the brake cylinder to change frequently, and generates a reaction force on the brake cylinder during the change, causing the cylinder bolts to deform and break. Therefore, the uneven wear of the brake pad will cause uneven force on the brake cylinder pressure column, which will affect the safe and stable operation of the wind turbine generator set and cause great difficulties for subsequent inspection and maintenance.

[0005] There is no effective solution to the problem of unstable operation of wind turbines caused by uneven wear of brake pads in the related art. Summary of the invention

[0006] The embodiments of the present application relate to a method and device for determining a wear state, a storage medium, and an electronic device. Through the embodiments of the present application, the problem of unstable operation of a wind turbine generator set caused by uneven wear of brake pads in the related art can be solved.

[0007] According to an embodiment of the present application, a method for determining a wear state is provided, comprising: obtaining first wear state data of a first brake pad, wherein the first wear state data comprises at least: a vibration signal and a temperature signal corresponding to the first brake pad; determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad; inputting the first wear state feature vector into a trained target diagnosis model, so that the trained target diagnosis model determines a target wear state of the first brake pad according to the first wear state feature vector.

[0008] In an exemplary embodiment, determining a multi-frequency signal characteristic corresponding to the first brake pad according to the first wear state data includes: performing frequency domain analysis and time domain analysis on the first wear state data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear state data, and performing wavelet packet analysis on the first wear state data to obtain wavelet packet energy corresponding to the first wear state data, wherein the multi-frequency signal characteristics include: the time domain distribution characteristics, the frequency domain distribution characteristics and the wavelet packet energy.

[0009] In an exemplary embodiment, the first wear status data is subjected to frequency domain analysis and time domain analysis to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear status data, including: preprocessing the first wear status data to remove noise in the first wear status data; determining waveform characteristics corresponding to the preprocessed first wear status data on a time series corresponding to the preprocessed first wear status data; and converting the preprocessed first wear status data into the frequency domain and determining the spectral characteristics of the preprocessed first wear status data in the frequency domain; determining the time domain distribution characteristics based on the waveform characteristics, and determining the frequency domain distribution characteristics based on the spectral characteristics.

[0010] In an exemplary embodiment, wavelet packet analysis is performed on the first wear state data to obtain wavelet packet energy corresponding to the first wear state data, including: preprocessing the first wear state data to remove noise in the first wear state data, and determining a target wavelet basis function based on the preprocessed first wear state data; decomposing the preprocessed first wear state data based on the target wavelet basis function to decompose the preprocessed first wear state data into multiple sub-bands, and determining the wavelet packet coefficient corresponding to each sub-band; and performing feature extraction on each wavelet packet coefficient to determine the wavelet packet energy.

[0011] In an exemplary embodiment, mapping the multi-frequency signal features to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad includes: mapping the multi-frequency signal features to the high-dimensional feature space through nonlinear mapping; performing feature extraction based on linear operations on the multi-frequency signal features in the high-dimensional feature space to obtain the first wear state feature vector.

[0012] In an exemplary embodiment, before inputting the first wear state feature vector into the trained target diagnosis model, the method further includes: obtaining second wear state data corresponding to the second brake pad, wherein the second wear state data at least includes: a vibration signal and a temperature signal corresponding to the second brake pad; determining a second wear state feature vector corresponding to the second brake pad based on the second wear state data; and training the target diagnosis model based on the second wear state feature vector to generate the trained target diagnosis model.

[0013] In an exemplary embodiment, the target diagnostic model is trained according to the second wear state feature vector to generate the trained target diagnostic model, including: determining the target network structure corresponding to the target diagnostic model according to the dimension of the second wear state feature vector; inputting the second wear state feature vector into the target diagnostic model including the target network structure, and determining multiple weak classifiers of the target diagnostic model according to the second wear state feature vector; determining the weight of each weak classifier according to the error rate of each weak classifier, and merging the multiple weak classifiers into a strong classifier through the target algorithm and the weight to generate the trained target diagnostic model.

[0014] According to another embodiment of the embodiment of the present application, a wear state determination device is also provided, including: an acquisition module, used to obtain first wear state data of a first brake pad, wherein the first wear state data includes at least: a vibration signal and a temperature signal corresponding to the first brake pad; a determination module, used to determine the multi-frequency signal characteristics corresponding to the first brake pad according to the first wear state data, and map the multi-frequency signal characteristics to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad; a diagnosis module, used to input the first wear state feature vector into a trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0015] According to another aspect of the embodiments of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the above method when running.

[0016] According to another aspect of the embodiments of the present application, an electronic device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the method through the computer program.

[0017] According to another aspect of the embodiments of the present application, a computer program product is provided, including a computer program, wherein the computer program is executed by a processor to perform the method.

[0018] In the embodiment of the present application, the first wear state data of the first brake pad, which includes at least a vibration signal and a temperature signal, is obtained; the multi-frequency signal characteristics corresponding to the first brake pad are determined according to the first wear state data, and the high-frequency signal characteristics are mapped to the high-order feature space to obtain the first wear state feature vector corresponding to the first brake pad; the first wear state feature vector is input into the trained target diagnosis model so that the trained target diagnosis model can diagnose the first brake pad and then determine the target wear state of the first brake pad. That is to say, the embodiment of the present application obtains the first wear state data of the first brake pad, determines the first wear state feature vector according to the first wear state data, and then inputs the first wear state data into the trained target diagnosis model to determine the target wear state of the first brake pad according to the output result of the trained target diagnosis model. After determining the target wear state of the target brake pad, corresponding preventive measures can be taken according to the target wear state of the target brake pad to avoid the problem of unstable operation of the wind turbine generator set caused by uneven wear of the brake pad, and then timely discover the wear state of the brake pad so that the wind turbine generator set can operate safely and stably. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0020] Figure 1 It is a hardware structure block diagram of a computer terminal of a method for determining a wear state according to an embodiment of the present application;

[0021] Figure 2 is a flow chart of a method for determining a wear state according to an embodiment of the present application;

[0022] Figure 3 is a schematic diagram of a wind turbine yaw brake wear capacitance displacement sensor according to an optional embodiment of the present application;

[0023] Figure 4 It is a structural block diagram of a device for determining a wear state according to an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0025] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products, or devices; "plurality" means two or more.

[0026] The method embodiments provided in the embodiments of the present application can be executed in a computer terminal or similar computing device or cloud platform or independent physical server or software platform, wherein the above software platform is run through one or more servers. Taking running on a computer terminal as an example, Figure 1 1 is a hardware structure block diagram of a computer terminal of a method for determining a wear state according to an embodiment of the present application. Figure 1 As shown, the computer terminal may include one or more ( Figure 1 Only one is shown in the figure) processor 102 and memory 104 for storing data. In an exemplary embodiment, the computer terminal may also include a transmission device 106 for communication functions and an input and output device 108, wherein the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA. A person skilled in the art can understand that Figure 1 The structure shown is only for illustration and does not limit the structure of the above-mentioned computer terminal. Figure 1 More or fewer components as shown, or with Figure 1 Equivalent functions or comparisons shown Figure 1 A different configuration with more features is shown.

[0027] The memory 104 can be used to store computer programs, for example, software programs and modules of application software, such as computer programs corresponding to the methods in the embodiments of the present application. The processor 102 executes various functional applications and data processing by running the computer programs stored in the memory 104, that is, the above-mentioned methods are implemented. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some examples, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0028] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of a computer terminal. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, referred to as NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0029] In this embodiment, a method for determining a wear state is provided, which is applied to the above-mentioned computer terminal. Figure 2 is a flow chart of a method for determining a wear state according to an embodiment of the present application, the flow chart includes the following steps S202 to S206:

[0030] Step S202, obtaining first wear state data of the first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad;

[0031] The first brake pad may be a brake pad in a wind turbine generator set, and a capacitive displacement sensor and a thermometer may be installed on the brake disc of the wind turbine generator set. The vibration signal corresponding to the first brake pad may be determined by the capacitive displacement sensor, and the temperature signal corresponding to the first brake pad may be determined by the thermometer.

[0032] Capacitive displacement sensors have high measurement accuracy (up to 7pm resolution) and can provide more accurate measurements. Using capacitive displacement sensors, the thickness of the brake can be accurately measured, allowing for better evaluation of the performance and service life of the brake. The fast response speed of the capacitive displacement sensor (up to 10Khz) enables real-time monitoring of the thickness of the brake.

[0033] The above-mentioned capacitive sensor and temperature meter are only examples, and in actual applications, other instruments may be used to determine the vibration signal and temperature signal corresponding to the first brake pad.

[0034] Step S204, determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad;

[0035] Step S206: input the first wear state feature vector into the trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0036] In the embodiment of the present application, the first wear state data of the first brake pad including at least a vibration signal and a temperature signal is obtained; the multi-frequency signal characteristics corresponding to the first brake pad are determined according to the first wear state data, and the high-frequency signal characteristics are mapped to the high-order feature space to obtain the first wear state feature vector corresponding to the first brake pad; the first wear state feature vector is input into the trained target diagnosis model so that the trained target diagnosis model diagnoses the first brake pad and then determines the target wear state of the first brake pad. That is to say, the embodiment of the present application obtains the first wear state data of the first brake pad, determines the first wear state feature vector according to the first wear state data, and then inputs the first wear state data into the trained target diagnosis model to determine the target wear state of the first brake pad according to the output result of the trained target diagnosis model. After determining the target wear state of the target brake pad, corresponding preventive measures can be taken according to the target wear state of the target brake pad to avoid the problem of unstable operation of the wind turbine generator set caused by uneven wear of the brake pad, and then timely discover the wear state of the brake pad to ensure safe and stable operation of the wind turbine generator set.

[0037] The optional step S204 of determining the multi-frequency signal characteristics corresponding to the first brake pad based on the first wear status data includes: performing frequency domain analysis and time domain analysis on the first wear status data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear status data, and performing wavelet packet analysis on the first wear status data to obtain wavelet packet energy corresponding to the first wear status data, wherein the multi-frequency signal characteristics include: the time domain distribution characteristics, the frequency domain distribution characteristics and the wavelet packet energy.

[0038] It can be understood that the determination of the multi-frequency signal characteristics of the wear state data of the first brake pad mainly includes frequency domain analysis, time domain analysis and wavelet packet analysis. These three analysis methods extract the characteristics of the brake pad wear state from different angles, which helps to more comprehensively understand the health status of the brake pad.

[0039] 1) Frequency domain analysis.

[0040] Frequency domain analysis involves converting a signal from the time domain to the frequency domain to identify different frequency components in the signal. For wear status data, this may include vibration signals, sound signals, or other frequency-related signals. By performing frequency domain analysis, specific frequency vibration patterns during the wear of the brake pad can be detected, which may be directly related to the degree of wear of the brake pad.

[0041] For example, suppose a vibration signal is collected from the yaw brake of a wind turbine. By performing a fast Fourier transform (FFT) on this signal, a spectrum can be obtained, showing the vibration intensity at different frequencies. If it is found that the vibration intensity at a certain frequency increases significantly when the brake pad is worn, then the vibration intensity at this frequency can be used as one of the frequency domain distribution characteristics of brake pad wear.

[0042] 2) Time domain analysis.

[0043] Time domain analysis focuses on the performance of the signal on the time axis. By analyzing the time domain characteristics of the signal, we can understand the duration, amplitude, time interval and other characteristics of the vibration during the wear of the brake pad. These characteristics are very important for understanding the dynamic process of brake pad wear.

[0044] For example, in time domain analysis, statistical features such as the root mean square (RMS), peak value, kurtosis, and skewness of the vibration signal may be calculated. For example, if the brake pad is found to be gradually worn, the root mean square value and peak value of the vibration signal will increase accordingly, indicating that the wear of the brake pad is related to the increase in vibration intensity in the time domain, and these features can quantify this change.

[0045] 3) Wavelet packet analysis.

[0046] Wavelet packet analysis is a more sophisticated time-frequency analysis technique that can decompose the signal into multiple sub-bands with different frequencies and time resolutions. This method is particularly suitable for the analysis of non-stationary signals and can reveal the change patterns of different frequency components over time during the wear of brake pads.

[0047] For example, in wavelet packet analysis, we first select a suitable wavelet function (wavelet basis) to perform multi-layer decomposition on the vibration signal. Each layer of decomposition will produce a set of approximate coefficients and detail coefficients, which reflect the low-frequency and high-frequency characteristics of the signal. By calculating the energy of the detail coefficients of each layer, we can get the wavelet packet energy, which is a comprehensive feature reflecting the degree of wear of the brake pad.

[0048] In practical applications, such as brake pad wear monitoring for wind turbines, the above three analyses can be performed through the collected vibration signals, temperature signals and other related signals to obtain the following characteristics:

[0049] Frequency domain characteristics: including vibration intensity and spectrum density at specific frequencies.

[0050] Time domain characteristics: including the root mean square value, peak value, kurtosis, skewness and other statistical characteristics of the vibration signal.

[0051] Wavelet packet energy: Calculated from the detail coefficients of each layer of wavelet packet decomposition, it reflects the comprehensive energy distribution of the signal in different sub-bands.

[0052] These features are then input into the BP_AdaBoost diagnostic model, which is trained to learn the relationship between these features and the wear status of the brake pads, so that it can identify whether the brake pads are in a normal, slightly worn or severely worn state based on the real-time monitored signal features. In this way, real-time and accurate monitoring of the wear status of the brake pads can be achieved, which helps prevent the failure of wind turbines and ensure their safe and stable operation.

[0053] That is, to determine the multi-frequency signal characteristics corresponding to the first brake pad, it is necessary to determine the time domain distribution characteristics, frequency domain distribution characteristics and wavelet packet energy. Specifically:

[0054] (1) performing frequency domain analysis and time domain analysis on the first wear status data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear status data, including: preprocessing the first wear status data to remove noise in the first wear status data; determining waveform characteristics corresponding to the preprocessed first wear status data on the time series corresponding to the preprocessed first wear status data; and converting the preprocessed first wear status data into the frequency domain and determining the spectrum characteristics of the preprocessed first wear status data in the frequency domain; determining the time domain distribution characteristics according to the waveform characteristics, and determining the frequency domain distribution characteristics according to the spectrum characteristics.

[0055] It can be understood that time domain analysis and frequency domain analysis mainly include data preprocessing, determining time domain waveform features, extracting frequency domain spectrum features, and analysis based on these features. Specifically:

[0056] Data preprocessing is an important step before analysis, which aims to remove noise from the data to make subsequent feature extraction more accurate. Noise may be introduced by sensor instability, environmental interference, or errors in signal transmission.

[0057] For example, assume that the vibration signal collected from the wind turbine brake contains noise, which will interfere with the identification of the wear state of the brake pad. The preprocessing process may include applying a low-pass filter to remove high-frequency noise, or using wavelet transform for denoising. For example, wavelets can be used for multi-layer decomposition, and then in each sub-band, noise can be selectively removed based on the ratio of signal energy to noise energy.

[0058] Determine the time domain waveform characteristics: The time domain waveform characteristics reflect the performance of the signal in the time series, including amplitude, waveform change rate, duration, etc.

[0059] For example, in the time domain, you can calculate the root mean square value (RMS), peak amplitude, waveform factor, kurtosis, and skewness of the vibration signal. For example, if the brake pads are worn, the root mean square value and peak amplitude of the vibration signal may increase because wear increases the vibration during braking.

[0060] Frequency Domain Spectral Characterization: Frequency Domain Spectral Characterization involves converting the signal into the frequency domain to identify different frequency components in the signal. It helps to discover specific frequency patterns associated with brake pad wear.

[0061] For example, by using Fast Fourier Transform (FFT) to convert the vibration signal in the time domain to the frequency domain, a spectrum can be obtained, which shows the vibration energy at different frequencies. If the brake pads are worn, some new frequency peaks may appear in the spectrum, or the vibration energy of certain frequencies may increase significantly. The vibration energy of these specific frequencies can be used as frequency domain distribution features.

[0062] Determine time domain and frequency domain distribution features: Determine time domain distribution features and frequency domain distribution features based on waveform features and spectrum features. These can be statistics calculated directly from preprocessed data or more complex features obtained through further analysis (e.g., wavelet packet analysis).

[0063] For example, suppose the root mean square value, form factor, specific frequency vibration energy and vibration energy ratio of the vibration signal are calculated. These values ​​can be used as distribution features in the time domain and frequency domain for subsequent BP_AdaBoost algorithm training.

[0064] Data preprocessing may also include signal normalization, detrending, etc., to make the signal characteristics comparable at different time points or between different sensors. When determining the time domain and frequency domain distribution characteristics, in addition to the basic characteristics mentioned above, more complex characteristics such as instantaneous frequency, modulation frequency, frequency bandwidth, etc. can also be considered to more comprehensively describe the dynamic characteristics of brake pad wear.

[0065] (2) performing wavelet packet analysis on the first wear state data to obtain wavelet packet energy corresponding to the first wear state data, including: preprocessing the first wear state data to remove noise in the first wear state data, and determining a target wavelet basis function based on the preprocessed first wear state data; decomposing the preprocessed first wear state data based on the target wavelet basis function to decompose the preprocessed first wear state data into a plurality of sub-bands, and determining a wavelet packet coefficient corresponding to each sub-band; and performing feature extraction on each wavelet packet coefficient to determine the wavelet packet energy.

[0066] It is understood that wavelet packet analysis includes:

[0067] Data preprocessing and noise reduction:

[0068] Data preprocessing usually includes steps such as signal denoising, detrending, and normalization to ensure the quality of the signal and make it more suitable for subsequent feature extraction and analysis. Denoising is a key step, which aims to remove interference components in the signal that are not related to the wear state, such as environmental noise, sensor noise, etc., to improve the clarity and signal-to-noise ratio of the signal.

[0069] For example, suppose a vibration signal is collected from the brake system of a wind turbine, but due to the ambient wind noise and sensor noise, the signal contains a large number of irrelevant high-frequency components. In the preprocessing stage, these high-frequency noises can be filtered out by applying a bandpass filter or wavelet denoising method, retaining the frequency components related to the wear status of the brake pads.

[0070] Determine the target wavelet basis function:

[0071] The selection of wavelet basis function is crucial for subsequent feature extraction. Different wavelet basis functions will have different performances in the signal decomposition process, so it is necessary to determine the most suitable wavelet basis function based on the characteristics of the preprocessed data.

[0072] For example, based on the preprocessed vibration signal, it may be found that the signal contains a large number of low-frequency and medium-frequency components, but fewer high-frequency components. In this case, a wavelet basis function with better low-frequency and medium-frequency decomposition performance can be selected.

[0073] Wavelet packet decomposition based on the target wavelet basis function:

[0074] The preprocessed data is decomposed into wavelet packets through the selected wavelet basis function, which can decompose the original signal into multiple sub-band signals with different frequencies and time resolutions. Each sub-band signal has corresponding wavelet packet coefficients, which reflect the local changes of the signal within the frequency range.

[0075] For example, if the frequency range of the vibration signal covers 0Hz to 1000Hz, it can be further subdivided into multiple sub-bands using wavelet packet decomposition, such as 0-125Hz, 125-250Hz, 250-500Hz, 500-1000Hz. The wavelet packet coefficients of each sub-band contain detailed information about the signal within the frequency range, which is extremely important for analyzing the wear pattern of the brake pad.

[0076] Feature extraction of wavelet packet energy:

[0077] The wavelet packet energy is obtained by calculating the absolute value square sum of the wavelet packet coefficients in each sub-band. It reflects the comprehensive energy distribution of the signal in different sub-bands and can be used to quantify the change in the degree of brake pad wear.

[0078] For example, the wavelet packet coefficients of each subband are calculated to obtain the wavelet packet energy. If the wavelet packet energy increases significantly within a certain frequency range, this may indicate that the brake pad begins to wear abnormally at this frequency. The trend of wavelet packet energy changes can be used as one of the input features of the diagnostic model to help the BP_AdaBoost algorithm identify the wear status of the brake pad.

[0079] Through the above steps, features that are highly correlated with the brake pad wear status, such as wavelet packet energy, can be extracted from the vibration signal. These features are then input into the BP_AdaBoost diagnostic model for machine learning training to achieve intelligent identification and prediction of the brake pad wear status.

[0080] Optionally, the above-mentioned step S204 of mapping the multi-frequency signal features to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad includes: mapping the multi-frequency signal features to the high-dimensional feature space through nonlinear mapping; and performing feature extraction based on linear operations on the multi-frequency signal features in the high-dimensional feature space to obtain the first wear state feature vector.

[0081] It is understandable that mapping the "multi-frequency signal features" to the "high-dimensional feature space" and obtaining the "first wear state feature vector" is a key step in the fusion application of kernel principal component analysis and BP_AdaBoost algorithm. This process helps to extract more distinguishing features from the original signal, thereby improving the model's recognition accuracy of the brake pad wear state. Specifically:

[0082] Nonlinear mapping to high-dimensional feature space:

[0083] Nonlinear mapping refers to converting the original low-dimensional features (i.e., features extracted by time domain, frequency domain, and wavelet packet analysis) into a high-dimensional feature space through a series of nonlinear functions. This conversion process can capture the nonlinear relationships between features, which may not be obvious in the original low-dimensional space.

[0084] For example, suppose that a vibration signal of brake pad wear is collected from the yaw brake of a wind turbine, and a series of features are obtained through time domain analysis, frequency domain analysis, and wavelet packet analysis, including the root mean square value of the vibration signal, frequency peak, wavelet packet energy distribution, etc. Based on these features, a nonlinear mapping (for example, Gaussian radial basis function (RBF) kernel function) is used to transform these features into a higher dimensional space.

[0085] Feature extraction in high-dimensional feature space:

[0086] Feature extraction in high-dimensional feature space usually refers to applying linear methods (such as principal component analysis) in the transformed space to further extract features. Since the original features become more "linearly separable" after nonlinear mapping, even linear methods can effectively extract features with strong classification capabilities in high-dimensional space.

[0087] For example: extracting features in a high-dimensional feature space. This is achieved by calculating the kernel matrix, which is the inner product matrix of the original features after nonlinear mapping. By solving the eigenvalues ​​and eigenvectors of the kernel matrix, the principal component that can best explain the data variation can be found.

[0088] It is also possible to obtain a feature vector containing multiple principal components, which represent the directions that contribute most to data changes in high-dimensional space. The first few principal components with the greatest explanatory power can be selected to form a new feature vector, which contains key information about the wear status of the brake pad.

[0089] The wear state feature vector contains a high-dimensional representation of the brake pad wear state, which can effectively express the characteristics of the brake pad at different wear levels. In the BP_AdaBoost algorithm, these feature vectors are used as input to train the neural network classifier to achieve accurate identification of the brake pad wear state.

[0090] Through the above process, the multi-frequency information in the original signal can be fully utilized, and a more complex and more discriminative feature vector can be obtained through nonlinear mapping and feature extraction in high-dimensional feature space. This helps to improve the performance and accuracy of the BP_AdaBoost algorithm in identifying the wear status of brake pads, thereby improving the safety and reliability of the wind turbine braking system.

[0091] Optionally, before inputting the first wear state feature vector into the trained target diagnosis model in the above step S208, the method also includes: obtaining second wear state data corresponding to the second brake pad, wherein the second wear state data at least includes: a vibration signal and a temperature signal corresponding to the second brake pad; determining a second wear state feature vector corresponding to the second brake pad based on the second wear state data; and training the target diagnosis model based on the second wear state feature vector to generate the trained target diagnosis model.

[0092] Among them, the target diagnosis model is trained according to the second wear state feature vector to generate the trained target diagnosis model, including: determining the target network structure corresponding to the target diagnosis model according to the dimension of the second wear state feature vector; inputting the second wear state feature vector into the target diagnosis model including the target network structure, and determining multiple weak classifiers of the target diagnosis model according to the second wear state feature vector; determining the weight of each weak classifier according to the error rate of each weak classifier, and merging the multiple weak classifiers into a strong classifier through the target algorithm and the weight to generate the trained target diagnosis model.

[0093] It is understandable that the use of BP_AdaBoost algorithm to build a diagnostic model for brake pad wear involves key steps such as determining the network structure based on the feature vector dimension, training multiple weak classifiers, determining the weights of weak classifiers, and merging weak classifiers to generate strong classifiers. Specifically:

[0094] Determine the network structure of the target diagnosis model: Determine the target network structure of the BP neural network according to the dimension of the second wear state feature vector. This is because the number of nodes in the input layer of the BP neural network needs to match the dimension of the feature vector to ensure that the model can correctly receive and process all input features.

[0095] For example: if the second wear state feature vector obtained by kernel principal component analysis (KPCA) contains 10 principal components (ie, the dimension of the feature vector is 10), then the input layer of the BP neural network will be designed to consist of 10 nodes to receive the input of these 10 feature vectors.

[0096] Train multiple weak classifiers:

[0097] Use the BP algorithm to train multiple weak classifiers. A weak classifier refers to a model with weak classification ability alone, but can improve the overall classification performance through ensemble learning in the entire model. Each weak classifier is a BP neural network.

[0098] For example, based on the feature vector, 10 BP neural networks are trained as weak classifiers. These neural networks all use the same input features (i.e., the second wear state feature vector), but their hidden layer structures, weight initialization, and weighting of training data may be different, which makes each neural network show different capabilities in the classification task.

[0099] Determine the weights of the weak classifiers:

[0100] The weight of each weak classifier is determined by its classification error rate. In the BP_AdaBoost algorithm, weak classifiers with poor classification results will be assigned lower weights, while weak classifiers with better classification results will be assigned higher weights. In this way, the model will rely more on weak classifiers with better performance in the final classification.

[0101] For example, during the training process, the third weak classifier (BP neural network) may have the lowest classification error rate, which means it performs best in identifying the wear status of the brake pad. Therefore, the BP_AdaBoost algorithm will assign a higher weight to the third weak classifier, while the weights of other weak classifiers are relatively low.

[0102] Merge weak classifiers to generate strong classifiers:

[0103] Through the BP_AdaBoost algorithm and the determined weights, multiple weak classifiers are merged into a strong classifier, which has a higher classification accuracy than any single weak classifier. The generation of a strong classifier usually involves multiple iterations, each of which adjusts the weights of the weak classifiers and reweights the training data based on the current classification results to focus more on samples that are difficult to classify correctly.

[0104] For example: In each iteration, the BP_AdaBoost algorithm calculates the classification error rate of the weak classifier based on the classification results of the current weak classifier and the true label of the data point. Then, it adjusts the weight of the weak classifier according to this error rate. Suppose after multiple iterations, we have 10 weak classifiers, each with a corresponding weight. In the final fusion stage, the BP_AdaBoost algorithm will weight and merge the outputs of these 10 weak classifiers according to these weights to obtain the final strong classifier output, that is, the trained target diagnosis model.

[0105] The training process of the BP_AdaBoost algorithm is a dynamic adjustment process. It adjusts the training data of the next stage according to the classification results of the current stage, so that it pays more attention to those misclassified samples, and gradually improves the overall performance of the classifier in this way. In practical applications, the training and merging of weak classifiers can be based on the principle of gradient boosting to ensure the stability and robustness of the model in classification tasks.

[0106] In order to better understand the process of the above-mentioned wear state determination method, the above-mentioned wear state determination method is described below in combination with an optional embodiment, but is not used to limit the technical solution of the embodiment of the present application.

[0107] The brake pad detection device of the yaw brake system of the wind turbine is composed of the digital input points of the programmable logic controller (PLC) control system. The travel switch base and micro travel switch for detection are installed on the brake pad as the signal collection point of the brake pad wear. The detection signal is collected through the input point of the PLC to realize the wear detection of the brake pad. The PLC analyzes the detected brake pad state. As the brake pad wears, the thickness of the brake pad will continue to decrease. When the brake pad wears to the recognition value of the wear state by the back propagation_adaptive enhancement (BP_AdaBoost) algorithm, the micro travel switch is activated, the PLC input point detects the brake pad wear signal, the PLC control module sends a brake pad wear fault signal, and the wind turbine fails and shuts down.

[0108] Figure 3 Schematic diagram of a wind turbine yaw brake wear capacitance displacement sensor according to an optional embodiment of the present application, such as Figure 3 As shown, an optional embodiment of the present application proposes an online monitoring method for wind turbine brake wear based on the back propagation-adaptive enhancement (BP_AdaBoost) algorithm, which uses a sensor to measure the displacement of brake disc wear based on the capacitance principle. It consists of two electrode plates and a medium. When an object moves, the capacitance between the two electrode plates will change. By detecting this change, the displacement of the object can be accurately measured. The capacitive displacement sensor is used to detect the wear of the friction plate of the yaw brake of the wind turbine.

[0109] One of the electrodes is fixed on the brake caliper, and the other variable electrode is an elastic diaphragm, which is fixed on the brake disc. When the brake disc is worn, the elastic diaphragm is displaced to change the capacitance. This sensor can convert the displacement into an electrical signal and transmit it to the main control system. The operation and maintenance personnel can view the degree of wear of the brake disc in real time, set the warning position and the shutdown position, and issue a warning and shutdown signal when the friction pad is worn to the specified limit. The brake pad wear state monitoring method using the BP_AdaBoost algorithm includes the following steps:

[0110] Step 1: Acquisition of brake pad wear state data (ie, wear state data).

[0111] First, a capacitive displacement sensor and a temperature meter are installed on the brake disc to collect data on the wear status of the brake pad, and then the wear status signal is recorded, for example: the vibration signal and temperature signal of the brake pad are collected.

[0112] Step 2: Feature extraction of tool wear status.

[0113] (1) Feature extraction of vibration signals and temperature signals: Perform time domain, frequency domain and wavelet packet analysis on the vibration signals and temperature signals to extract their feature vectors (i.e., multi-frequency signal features).

[0114] (2) Kernel principal component analysis: For the energy decomposed by wavelet packets and the preprocessed frequency domain and time domain signals, they map the independent variables to a high-dimensional feature space through nonlinear mapping, and perform feature extraction based on linear operations in the high-dimensional feature space to form a new feature vector (i.e., the wear state feature vector).

[0115] Step 3: Train the BP_AdaBoost diagnostic model (i.e., the target diagnostic model).

[0116] The new feature vector extracted by kernel principal component analysis is used as the input of the BP_AdaBoost diagnostic model to train the BP_AdaBoost diagnostic model. According to the feature vector characteristics of the end mill wear state, the structure of the BP neural network is determined, where the input layer, hidden layer, and output layer are 15 / 10 / 3 respectively, the weights and thresholds of the neural network are randomly initialized, and a total of 10 BP neural network weak classifiers are trained and generated. Finally, the 10 weak classifiers form a strong classifier to recognize the wear state of the brake pad.

[0117] According to an optional embodiment of the present application, the capacitive displacement sensor has high measurement accuracy (up to 7pm resolution) and can provide more accurate measurements. Using the capacitive displacement sensor, we can accurately measure the thickness of the brake, so as to better evaluate the performance and service life of the brake; at the same time, the fast response speed of the capacitive displacement sensor (up to 10Khz) enables us to monitor the thickness of the brake in real time; in addition, the structure is simple, easy to install and use, and the BP_AdaBoost algorithm is used to identify the wear state of the brake pad, and the wear identification accuracy rate is as high as 90%.

[0118] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods of each embodiment of the present application.

[0119] The present application also provides a structural block diagram of a wear state determination device. Figure 4 is a structural block diagram of a device for determining a wear state according to an embodiment of the present application; Figure 4 As shown, including:

[0120] An acquisition module 42 is used to acquire first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad;

[0121] a determination module 44, configured to determine a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and map the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad;

[0122] The diagnosis module 46 is used to input the first wear state feature vector into the trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0123] Through the above device, the first wear state data of the first brake pad, which includes at least a vibration signal and a temperature signal, is obtained; the multi-frequency signal characteristics corresponding to the first brake pad are determined according to the first wear state data, and the high-frequency signal characteristics are mapped to the high-order feature space to obtain the first wear state feature vector corresponding to the first brake pad; the first wear state feature vector is input into the trained target diagnosis model so that the trained target diagnosis model can diagnose the first brake pad and then determine the target wear state of the first brake pad. That is to say, the embodiment of the present application obtains the first wear state data of the first brake pad, determines the first wear state feature vector according to the first wear state data, and then inputs the first wear state data into the trained target diagnosis model to determine the target wear state of the first brake pad according to the output result of the trained target diagnosis model. After determining the target wear state of the target brake pad, corresponding preventive measures can be taken according to the target wear state of the target brake pad to avoid the problem of unstable operation of the wind turbine generator set caused by uneven wear of the brake pad, and then timely discover the wear state of the brake pad to ensure safe and stable operation of the wind turbine generator set.

[0124] In an exemplary embodiment, the determination module 44 is further used to perform frequency domain analysis and time domain analysis on the first wear status data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear status data, and to perform wavelet packet analysis on the first wear status data to obtain wavelet packet energy corresponding to the first wear status data, wherein the multi-frequency signal characteristics include: the time domain distribution characteristics, the frequency domain distribution characteristics and the wavelet packet energy.

[0125] In an exemplary embodiment, the determination module 44 is also used to preprocess the first wear status data to remove noise in the first wear status data; determine the waveform characteristics corresponding to the preprocessed first wear status data on the time series corresponding to the preprocessed first wear status data; and convert the preprocessed first wear status data into the frequency domain and determine the spectral characteristics of the preprocessed first wear status data in the frequency domain; determine the time domain distribution characteristics based on the waveform characteristics, and determine the frequency domain distribution characteristics based on the spectral characteristics.

[0126] In an exemplary embodiment, the determination module 44 is also used to preprocess the first wear status data to remove noise in the first wear status data, and determine the target wavelet basis function based on the preprocessed first wear status data; decompose the preprocessed first wear status data based on the target wavelet basis function to decompose the preprocessed first wear status data into multiple sub-bands, and determine the wavelet packet coefficient corresponding to each sub-band; perform feature extraction on each wavelet packet coefficient to determine the wavelet packet energy.

[0127] In an exemplary embodiment, the determination module 44 is further used to map the multi-frequency signal features to the high-dimensional feature space through nonlinear mapping; and perform feature extraction based on linear operations on the multi-frequency signal features in the high-dimensional feature space to obtain the first wear state feature vector.

[0128] In an exemplary embodiment, the diagnostic module 46 is also used to obtain second wear state data corresponding to the second brake pad, wherein the second wear state data at least includes: a vibration signal and a temperature signal corresponding to the second brake pad; determining a second wear state feature vector corresponding to the second brake pad based on the second wear state data; and training a target diagnostic model based on the second wear state feature vector to generate the trained target diagnostic model.

[0129] In an exemplary embodiment, the diagnostic module 46 is also used to determine the target network structure corresponding to the target diagnostic model according to the dimension of the second wear state feature vector; input the second wear state feature vector into the target diagnostic model including the target network structure, and determine multiple weak classifiers of the target diagnostic model according to the second wear state feature vector; determine the weight of each weak classifier according to the error rate of each weak classifier, and merge the multiple weak classifiers into a strong classifier through the target algorithm and the weights to generate the trained target diagnostic model.

[0130] An embodiment of the present application further provides a storage medium, which includes a stored program, wherein the program executes any of the above methods when it is run.

[0131] Optionally, in this embodiment, the storage medium may be configured to store program codes for executing the following steps:

[0132] S1, obtaining first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad;

[0133] S2, determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad;

[0134] S3, inputting the first wear state feature vector into the trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0135] An embodiment of the present application further provides an electronic device, including a memory and a processor, wherein a computer program is stored in the memory, and the processor is configured to run the computer program to execute the steps in any one of the above method embodiments.

[0136] Optionally, the electronic device may further include a transmission device and an input / output device, wherein the transmission device is connected to the processor, and the input / output device is connected to the processor.

[0137] Optionally, in this embodiment, the processor may be configured to perform the following steps through a computer program:

[0138] S1, obtaining first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad;

[0139] S2, determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad;

[0140] S3, inputting the first wear state feature vector into the trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0141] An embodiment of the present application further provides a computer program product, including a computer program, which is used by a processor to execute the steps in any of the above method embodiments.

[0142] Optionally, in this embodiment, the computer program product may be executed by a processor to perform the following steps:

[0143] S1, obtaining first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad;

[0144] S2, determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad;

[0145] S3, inputting the first wear state feature vector into the trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

[0146] Optionally, in this embodiment, the above-mentioned storage medium may include but is not limited to: a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a mobile hard disk, a magnetic disk or an optical disk, and other media that can store program codes.

[0147] Optionally, the specific examples in this embodiment may refer to the examples described in the above embodiments and optional implementation modes, and this embodiment will not be described in detail here.

[0148] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, and optionally, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order from that herein, or they can be made into individual integrated circuit modules, or multiple modules or steps therein can be made into a single integrated circuit module for implementation. Thus, the present application is not limited to any specific combination of hardware and software.

[0149] The above description is only the preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A method for determining a wear state, characterized in that: include: Acquire first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad; Determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad; The first wear state feature vector is input into a trained target diagnosis model, so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

2. The method for determining the wear state according to claim 1, characterized in that: Determining a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data includes: Perform frequency domain analysis and time domain analysis on the first wear state data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear state data, and perform wavelet packet analysis on the first wear state data to obtain wavelet packet energy corresponding to the first wear state data, wherein the multi-frequency signal characteristics include: the time domain distribution characteristics, the frequency domain distribution characteristics and the wavelet packet energy.

3. The method for determining the wear state according to claim 2, characterized in that: Performing frequency domain analysis and time domain analysis on the first wear state data to obtain time domain distribution characteristics and frequency domain distribution characteristics corresponding to the first wear state data includes: Preprocessing the first wear state data to remove noise in the first wear state data; determining waveform features corresponding to the preprocessed first wear state data on a time series corresponding to the preprocessed first wear state data; and, Converting the preprocessed first wear state data into a frequency domain, and determining frequency spectrum characteristics of the preprocessed first wear state data in the frequency domain; The time domain distribution feature is determined according to the waveform feature, and the frequency domain distribution feature is determined according to the spectrum feature.

4. The method for determining the wear state according to claim 2, characterized in that: Performing wavelet packet analysis on the first wear state data to obtain wavelet packet energy corresponding to the first wear state data includes: Preprocessing the first wear state data to remove noise in the first wear state data, and determining a target wavelet basis function according to the preprocessed first wear state data; decomposing the preprocessed first wear state data based on the target wavelet basis function to decompose the preprocessed first wear state data into a plurality of sub-bands, and determining a wavelet packet coefficient corresponding to each sub-band; Feature extraction is performed on each wavelet packet coefficient to determine the wavelet packet energy.

5. The method for determining the wear state according to claim 1, characterized in that: Mapping the multi-frequency signal feature to a high-dimensional feature space to obtain a first wear state feature vector corresponding to the first brake pad includes: Mapping the multi-frequency signal features to the high-dimensional feature space through nonlinear mapping; A feature extraction based on linear operation is performed on the multi-frequency signal features in the high-dimensional feature space to obtain the first wear state feature vector.

6. The method for determining the wear state according to claim 1, characterized in that: Before inputting the first wear state feature vector into the trained target diagnosis model, the method further includes: Acquire second wear state data corresponding to the second brake pad, wherein the second wear state data at least includes: a vibration signal and a temperature signal corresponding to the second brake pad; Determine a second wear state feature vector corresponding to the second brake pad according to the second wear state data; The target diagnosis model is trained according to the second wear state feature vector to generate the trained target diagnosis model.

7. The method for determining the wear state according to claim 6, characterized in that: Training a target diagnosis model according to the second wear state feature vector to generate the trained target diagnosis model includes: Determining a target network structure corresponding to the target diagnosis model according to the dimension of the second wear state feature vector; Inputting the second wear state feature vector into a target diagnosis model including a target network structure, and determining a plurality of weak classifiers of the target diagnosis model according to the second wear state feature vector; The weight of each weak classifier is determined according to the error rate of each weak classifier, and the multiple weak classifiers are merged into a strong classifier through a target algorithm and the weight to generate the trained target diagnosis model.

8. A device for determining a wear state, characterized in that: include: an acquisition module, configured to acquire first wear state data of a first brake pad, wherein the first wear state data at least includes: a vibration signal and a temperature signal corresponding to the first brake pad; a determination module, configured to determine a multi-frequency signal feature corresponding to the first brake pad according to the first wear state data, and map the multi-frequency signal feature to a high-dimensional feature space to acquire a first wear state feature vector corresponding to the first brake pad; A diagnosis module is used to input the first wear state feature vector into a trained target diagnosis model so that the trained target diagnosis model determines the target wear state of the first brake pad according to the first wear state feature vector.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored program, wherein the program executes the method described in any one of claims 1 to 7 when executed.

10. An electronic device comprising a memory and a processor, characterized in that: A computer program is stored in the memory, and the processor is configured to execute the method according to any one of claims 1 to 7 through the computer program.

Images