A method for extracting soundprint features of circuit breaker mechanical faults

Abstract

The present invention provides a method for extracting soundprint features from mechanical faults of circuit breakers, belonging to the technical field of soundprints for mechanical faults of circuit breakers. The method comprises the following steps: collecting sound signals from the circuit breaker during normal operation and various mechanical fault conditions; filtering and denoising the collected sound signals to obtain preprocessed sound signals, and extracting a set of abnormal waveform segments; processing the preprocessed sound signals using the GPU's CUDA parallel processing stream and the CPU's input stream processing of the abnormal waveform segments; extracting time-frequency domain features from the preprocessed sound signals to obtain a first feature matrix; clustering the first feature matrix to obtain multiple cluster centers as typical sound features and constructing a second feature matrix; extracting features from the abnormal waveform segments to obtain a third feature matrix, and performing feature fusion and dimensionality reduction with the second feature matrix; and finally obtaining a mechanical fault soundprint feature matrix. This method solves the technical problem of low efficiency of traditional serial processing methods.

Description

Technical Field

[0001] The present invention belongs to the technical field of circuit breaker mechanical fault soundprints, and in particular relates to a circuit breaker mechanical fault soundprint feature extraction method. Background Art

[0002] In modern power systems, circuit breakers, as critical switching devices, have a direct impact on the stability and safety of power supply. With the continuous expansion of power grids and the continued growth of electricity demand, higher requirements are being placed on circuit breaker operating status monitoring and fault diagnosis. Traditional circuit breaker fault diagnosis methods rely primarily on regular maintenance and manual inspections, which are not only time-consuming and labor-intensive but also difficult to detect potential fault hazards in a timely manner. In recent years, fault diagnosis technology based on sound signal analysis has gradually become a research hotspot due to its advantages such as non-invasiveness, real-time performance, and cost-effectiveness.

[0003] Currently, circuit breaker sound signal analysis mainly focuses on time domain analysis, frequency domain analysis, and time-frequency analysis. Time domain analysis focuses on characteristics such as the amplitude and duration of the sound signal, but its ability to identify complex faults is limited. Frequency domain analysis extracts spectral features through methods such as Fourier transform, but it is difficult to reflect the time-varying characteristics of the signal. Although time-frequency analysis such as short-time Fourier transform and wavelet transform can provide time domain and frequency domain information simultaneously, they still have limitations when processing non-stationary signals. In addition, existing methods often only focus on a single type of feature and ignore the relationship between different features, resulting in insufficient accuracy and reliability of fault diagnosis.

[0004] In practical applications, circuit breaker acoustic signals often contain complex background noise and transient interference, posing significant challenges for feature extraction. Existing noise reduction methods often oversmoothe the signal, resulting in the loss of useful fault information. Furthermore, due to the diversity and complexity of circuit breaker mechanical faults, a single feature extraction method struggles to fully capture all aspects of the fault, compromising the comprehensiveness and accuracy of diagnosis.

[0005] Furthermore, as data volumes increase, traditional methods for extracting voiceprint features from circuit breaker mechanical faults, often using a serial approach and inefficiently, are no longer able to meet the demands of real-time diagnosis. Effectively leveraging the parallel processing capabilities of modern computing hardware to improve feature extraction efficiency has also become a pressing issue.

[0006] Therefore, there is an urgent need for a method that can comprehensively utilize multiple feature information to efficiently and accurately extract features from circuit breaker sound signals under complex backgrounds, so as to improve the accuracy and real-time performance of mechanical fault diagnosis. Summary of the Invention

[0007] In view of this, the present invention provides a method for extracting voiceprint features of mechanical faults of circuit breaker, which can solve the technical problem that traditional methods for extracting voiceprint features of mechanical faults of circuit breaker often use a serial method and are less efficient.

[0008] The present invention is achieved in that:

[0009] A first aspect of the present invention provides a method for extracting voiceprint features of a circuit breaker mechanical fault, comprising the following steps:

[0010] S10, continuously collect sound signals during normal operation of the circuit breaker and various mechanical fault conditions;

[0011] S20, filtering and denoising the sound signal to obtain a preprocessed sound signal and abnormal sudden change segments of the sound signal waveform excluded during the filtering and denoising process, which are recorded as an abnormal waveform segment set;

[0012] S30, starting the CUDA parallel processing flow of the GPU to receive the pre-processed sound signal; starting the input flow of the CPU to receive the abnormal waveform segment set;

[0013] S40, using a sliding window to perform short-time Fourier transform, Mel-frequency cepstral coefficient analysis, and wavelet packet decomposition on the preprocessed sound signal to obtain features of the preprocessed sound signal, which are recorded as a first feature matrix;

[0014] S50, clustering the first feature matrix to obtain multiple cluster centers for representing typical sound features;

[0015] S60, for each element in the cluster, replacing the element with the cluster center to obtain a second feature matrix;

[0016] S70, extracting features of the abnormal waveform segment set to obtain a third feature matrix, and aligning the second feature matrix and the third feature matrix based on the second feature matrix;

[0017] S80, fusing the second feature matrix and the third feature matrix using a preset feature fusion equation group, and reducing the dimension using principal component analysis to obtain a feature matrix after dimension reduction;

[0018] S90. Standardize the feature matrix after dimensionality reduction to finally obtain a mechanical fault voiceprint feature matrix.

[0019] Specifically, step S10 includes:

[0020] Step 101: Select a high-sensitivity microphone array as a sound collection device, and install the high-sensitivity microphone array at key locations around the circuit breaker;

[0021] Step 102: Set the sampling frequency of the high-sensitivity microphone array to 44.1 kHz, and the acquisition time to 5 seconds after each operation;

[0022] Step 103: Use a synchronous trigger device to ensure that the collection of the sound signal starts synchronously with the operation of the circuit breaker;

[0023] Step 104: collect samples multiple times under different operating conditions, including normal operating state, slight wear state, severe wear state, and mechanical jam state, and collect at least 100 samples for each state;

[0024] Step 105: Save the collected sound signal in WAV format;

[0025] Step 106: Establish a database to record metadata such as environmental conditions, circuit breaker status, and collection time for each collection.

[0026] The step S20 specifically includes:

[0027] Step 201: Filter the signal using a bandpass filter, where the passband range of the filter is set to 100 Hz to 10 kHz;

[0028] Step 202: Use Butterworth filter design with the order selected as 8th order;

[0029] Step 203: Use the wavelet threshold denoising method to perform denoising, select the db4 wavelet basis, and set the decomposition level to 5;

[0030] Step 204: The wavelet coefficients of each layer are processed using a soft threshold function, and the threshold is selected using the Stein unbiased risk estimation criterion;

[0031] Step 205: During the noise reduction process, detect the mutation segment of the signal, define a mutation segment detection function, and mark the signal segment as an abnormal mutation segment when the detection function exceeds a preset threshold;

[0032] Step 206: Record the start time, duration, and amplitude information of the abnormal mutation segment, exclude the abnormal mutation segment from the preprocessed signal, and save it separately as an abnormal waveform segment set.

[0033] Wherein, the step S30 specifically includes:

[0034] Step 301: Initialize the NVIDIA CUDA environment and set GPU device parameters, including setting the thread block size to 256 and the grid size to be dynamically adjusted according to the data volume;

[0035] Step 302: Create a CUDA stream object for managing parallel tasks on the GPU;

[0036] Step 303: Copy the pre-processed sound signal data from the host memory to the global memory of the GPU, and use the CUDAMemcpyAsync function to implement asynchronous data transmission;

[0037] Step 304: Create an input stream object on the CPU side for receiving the abnormal waveform segment set. The CPU input stream adopts a double buffering mechanism and sets the buffer size to 1 megabyte.

[0038] Step 305: Implement a synchronization mechanism, using condition variables and mutexes to coordinate the data processing progress of the GPU and CPU;

[0039] Step 306: Implementing a CUDA kernel function on the GPU side for parallel processing of operations such as short-time Fourier transform of the preprocessed signal;

[0040] Step 307: Implement a multi-threaded processing function on the CPU side to process feature extraction of abnormal waveform segment sets, and set the number of CPU threads to twice the number of system logic cores.

[0041] The step S40 specifically includes:

[0042] Step 401: Perform short-time Fourier transform, select Hanning window as the window function, set the window length to 1024 sampling points, and the overlap rate to 50%;

[0043] Step 402: Calculate the Mel-frequency cepstral coefficients by first converting the short-time Fourier transform result to the Mel-frequency scale using a 40-Mel filter bank.

[0044] Step 403: Take the logarithm of the Mel filter output and perform discrete cosine transform, retaining the first 13 coefficients as Mel frequency cepstral coefficient features;

[0045] Step 404: perform wavelet packet decomposition, select the db4 wavelet basis, set the decomposition layer to 5, and calculate the energy characteristics of each sub-band;

[0046] Step 405: Combine the short-time Fourier transform, the Mel-frequency cepstral coefficients, and the wavelet packet energy features into a feature matrix, referred to as the first feature matrix;

[0047] Step 406: Normalize the first feature matrix using the Z-score normalization method.

[0048] The step S50 specifically includes:

[0049] Step 501: clustering is performed using the improved K-means++ algorithm, and the optimal number of clusters K is determined using the elbow rule;

[0050] Step 502: Calculate the total intra-class sum of squares under different K values, and select the inflection point where the rate of decrease of the total intra-class sum of squares slows down significantly as the optimal K value;

[0051] Step 503: When initializing the cluster centers, the first center is randomly selected, and the probability of selecting subsequent centers is proportional to the square of the minimum distance from the selected center to the first center.

[0052] Step 504: To improve the stability of clustering, a strategy of averaging multiple runs is adopted, specifically 10 runs, and the result with the smallest sum of squares within the overall cluster is selected;

[0053] Step 505: In each iteration, calculate the distance from each sample to its nearest center, mark samples with a distance greater than 3 times the standard deviation as potential outliers, and give them a smaller weight when updating the cluster center;

[0054] Step 506: After clustering is completed, the representativeness score of each cluster center is calculated, and several cluster centers with the highest scores are selected as typical sound features.

[0055] The step S60 specifically includes:

[0056] Step 601: Calculate the distance from each sample point to each cluster center, using Euclidean distance for distance calculation;

[0057] Step 602: For each sample point, find the closest cluster center and replace the coordinates of the original sample point with the coordinates of the cluster center;

[0058] Step 603: Introduce a soft assignment mechanism, define soft assignment weights, and calculate the new coordinates of the sample point as the weighted sum of the coordinates of all cluster centers;

[0059] Step 604: perform principal component analysis on the replaced feature matrix to calculate the covariance matrix;

[0060] Step 605: Solve the characteristic equation to obtain the eigenvalue and eigenvector, and select the top k principal components whose cumulative contribution rate reaches 95%;

[0061] Step 606: Project the data into the subspace formed by the k principal components to obtain a second feature matrix;

[0062] Step 607: normalize the second characteristic matrix using the Min-Max normalization method.

[0063] The step S70 specifically includes:

[0064] Step 701: extract features from each waveform segment in the abnormal waveform segment set, and calculate short-time Fourier transform, Mel-frequency cepstral coefficients, and wavelet packet energy features using the same method as step S40;

[0065] Step 702: Adjust the calculation of the short-time Fourier transform and the Mel-frequency cepstral coefficients accordingly, reduce the window length to 512 sampling points, and increase the overlap rate to 75%;

[0066] Step 703: For wavelet packet decomposition, keep the 5-layer decomposition, but only select the energy features of the high-frequency sub-band to highlight the abnormal features;

[0067] Step 704: Combining the extracted features into a third feature matrix;

[0068] Step 705: Use the timestamp information of the abnormal waveform segment to map it to the time frame of the second characteristic matrix to perform time alignment;

[0069] Step 706: Use a linear interpolation method to upsample or downsample the third feature matrix so that its dimension is consistent with that of the second feature matrix, and perform feature dimension alignment;

[0070] Step 707: Normalize the third characteristic matrix using the same Min-Max normalization method as that used for the second characteristic matrix.

[0071] The step S80 specifically includes:

[0072] Step 801: Apply a preset feature fusion equation group to perform feature fusion, including time-frequency domain fusion, energy distribution fusion, phase spectrum fusion and abnormal feature fusion;

[0073] Step 802: Use the particle swarm optimization algorithm to optimize the weight coefficients in the fusion equation, with the population size set to 50 and the maximum number of iterations to 100;

[0074] Step 803: The optimization objective function is the ratio of the inter-class distance to the intra-class distance of the fused features;

[0075] Step 804: performing centralization processing on the fused feature matrix;

[0076] Step 805: Calculate the covariance matrix, solve the characteristic equation, and obtain the eigenvalues ​​and corresponding eigenvectors;

[0077] Step 806: Arrange the eigenvalues ​​in descending order and select the top k principal components whose cumulative contribution rate reaches 95%;

[0078] Step 807: Construct a projection matrix, whose columns are selected eigenvectors, and obtain a dimensionality-reduced eigenmatrix through matrix multiplication.

[0079] The step S90 specifically includes:

[0080] Step 901: Use the box plot method to identify outliers, and consider values ​​less than the first quartile minus 1.5 times the interquartile range or greater than the third quartile plus 1.5 times the interquartile range as potential outliers;

[0081] Step 902: Apply winsorization to the detected outliers, limiting the outliers to the range from the first quartile minus 1.5 times the interquartile range to the third quartile plus 1.5 times the interquartile range;

[0082] Step 903: Perform Z-score normalization on the processed feature matrix;

[0083] Step 904: Calculate the correlation matrix using the Pearson correlation coefficient to check the correlation between features;

[0084] Step 905: If there is a highly correlated feature pair with an absolute value of the correlation coefficient greater than 0.9, consider deleting one of the features to reduce the multicollinearity problem;

[0085] Step 906: Name and describe the standardized features, naming them according to their origin and physical meaning;

[0086] Step 907: Construct a feature importance ranking based on the variance contribution rate of the feature or the feature importance score of a model such as a random forest, and append the feature importance information to the final feature matrix.

[0087] The clustering is achieved by using an improved K-means++ algorithm.

[0088] Furthermore, 1. Time-frequency domain fusion equation:

[0089]

[0090] Among them: F t (x, y) is the fused time-frequency domain feature; S(x, y) is the time-frequency domain feature in the second feature matrix; T(x, y) is the time-frequency domain feature in the third feature matrix; α, β, γ, δ are weight coefficients obtained by particle swarm optimization algorithm; is the second-order partial derivative of S(x,y), which is used to enhance edge information; A(x,y) is the time-frequency feature of the abnormal mutation segment, which is obtained by short-time Fourier transform.

[0091] 2. Energy distribution fusion equation:

[0092]

[0093] Where: E(f) is the energy distribution characteristic after fusion; P i (f) is the power spectral density of the i-th feature matrix; w i is the weight coefficient, calculated by entropy weight method; N is the number of feature matrices (here 2); η is the weight coefficient of abnormal mutation segment energy; M is the number of abnormal mutation segments; D j is the duration of the jth abnormal mutation segment; T j is the total duration of the signal; E j (f) is the energy spectrum density of the jth abnormal mutation segment.

[0094] 3. Phase spectrum fusion equation:

[0095]

[0096] Where: Φ(f) is the phase spectrum feature after fusion; φ i (f) is the phase spectrum of the i-th characteristic matrix; v i is the weight coefficient, obtained through phase consistency evaluation; k is an integer used to ensure phase continuity; μ is the weight coefficient of the abnormal mutation phase; A j is the mutation amplitude of the jth abnormal mutation segment; A max is the maximum mutation amplitude among all abnormal mutation segments; φ j (f) is the phase spectrum of the jth abnormal mutation segment.

[0097] 4. Abnormal feature fusion equation:

[0098]

[0099] Among them: F a (t) is the abnormal feature after fusion; M is the number of abnormal mutation segments; λ j is the weight coefficient of the jth abnormal mutation segment, which is calculated comprehensively by the mutation amplitude and duration; t j,start and t j,end are the start and end time of the jth abnormal mutation segment; t j,mid is the midpoint of the jth abnormal mutation segment; σ j is the time scale parameter of the jth abnormal mutation segment, which is related to the duration; A j is the mutation amplitude of the jth abnormal mutation segment.

[0100] in,

[0101] Among them, s a (t) is the abnormal mutation segment signal, and w(t) is the window function (such as Hanning window).

[0102]

[0103] Among them, s i (t) is the time domain signal corresponding to the i-th feature matrix, and T is the signal duration.

[0104]

[0105] Where, Hi=-∑fpi(f)logpi(f),

[0106]

[0107] Among them, s j (t) is the signal of the jth abnormal mutation segment.

[0108]

[0109] in, is the average phase of all eigenmatrices at frequency f.

[0110]

[0111] Among them, X i (f) is the Fourier transform of the i-th characteristic matrix.

[0112]

[0113]

[0114] This setting makes the value of the Gaussian function at the beginning and end of the abnormal mutation segment approximately 1% of the peak value.

[0115]

[0116] D=tj,endd-tj,start

[0117]

[0118] in, is the local mean or trend of the signal.

[0119]

[0120] These calculation formulas provide a method for deriving the individual variables. In practical applications, appropriate adjustments and optimizations may be required based on specific signal characteristics and system requirements. For example, the weight coefficients (α, β, γ, δ, η, μ) can be optimized using machine learning algorithms (such as genetic algorithms or particle swarm optimization) to achieve optimal fusion results.

[0121] Compared with the prior art, the method for extracting circuit breaker mechanical fault voiceprint features provided by the present invention has the following beneficial effects:

[0122] First, the present invention employs a multimodal feature fusion method, comprehensively utilizing time-frequency domain features, energy distribution features, phase spectrum features, and anomaly features to comprehensively capture all aspects of the circuit breaker sound signal. This multi-dimensional feature extraction method greatly enhances the ability to identify complex mechanical faults, overcoming the limitations of single feature extraction methods.

[0123] Secondly, this invention introduces innovative preprocessing techniques. By combining bandpass filtering and wavelet threshold noise reduction, it effectively removes background noise while retaining critical fault information. In particular, the separate processing of abnormal and sudden changes ensures that transient fault information is not lost during the noise reduction process, greatly improving the accuracy of feature extraction.

[0124] Thirdly, this invention utilizes a GPU-CPU collaborative processing architecture, fully leveraging the parallel computing capabilities of modern hardware. The GPU handles large-scale data parallel processing, such as computationally intensive tasks like short-time Fourier transforms, while the CPU handles tasks requiring flexible control, such as abnormal waveform segments. This collaborative processing mechanism significantly improves feature extraction efficiency, making real-time monitoring of large-scale circuit breaker fleets possible.

[0125] Furthermore, this invention incorporates cluster analysis and a soft allocation mechanism to effectively reduce data redundancy and extract the most representative sound features. This not only reduces the computational complexity of subsequent processing but also improves the expressiveness of the features. Furthermore, dimensionality reduction through principal component analysis further optimizes the feature space, improving the efficiency and accuracy of subsequent fault diagnosis.

[0126] Finally, this invention employs an innovative set of feature fusion equations, including time-frequency domain fusion, energy distribution fusion, phase spectrum fusion, and anomaly feature fusion. This multi-layered fusion strategy leverages the strengths of various features, significantly improving their ability to distinguish. In particular, the introduction of the anomaly feature fusion equations effectively enhances the ability to identify transient faults.

[0127] In general, the method of the present invention solves the technical problem that traditional circuit breaker mechanical fault voiceprint feature extraction methods often use a serial approach and are less efficient. BRIEF DESCRIPTION OF THE DRAWINGS

[0128] Figure 1 A flow chart of the method provided by the present invention. DETAILED DESCRIPTION

[0129] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention.

[0130] like Figure 1 FIG. 1 is a flow chart of a method for extracting voiceprint features of a circuit breaker mechanical fault provided by the present invention. The method includes the following steps:

[0131] S10, continuously collect sound signals during normal operation of the circuit breaker and various mechanical fault conditions;

[0132] S20, filtering and denoising the sound signal to obtain a preprocessed sound signal and abnormal sudden change segments of the sound signal waveform excluded in the filtering and denoising process, which are recorded as an abnormal waveform segment set;

[0133] S30, starting the CUDA parallel processing flow of the GPU to receive the preprocessed sound signal; starting the input flow of the CPU to receive the abnormal waveform segment set;

[0134] S40, using a sliding window to perform short-time Fourier transform, Mel-frequency cepstral coefficient analysis, and wavelet packet decomposition on the preprocessed sound signal to obtain features of the preprocessed sound signal, which are recorded as a first feature matrix;

[0135] S50, clustering the first feature matrix to obtain multiple cluster centers for representing typical sound features;

[0136] S60, for each element in the cluster, using the cluster center to replace the element to obtain a second feature matrix;

[0137] S70, extracting features of the abnormal waveform segment set to obtain a third feature matrix, and aligning the second feature matrix and the third feature matrix based on the second feature matrix;

[0138] S80, fusing the second characteristic matrix and the third characteristic matrix using a preset characteristic fusion equation group, and reducing the dimension using a principal component analysis method to obtain a characteristic matrix after dimension reduction;

[0139] S90. Standardize the feature matrix after dimensionality reduction to finally obtain a mechanical fault voiceprint feature matrix.

[0140] The specific implementation of the above steps is described in detail below:

[0141] The specific implementation of step S10 is as follows: First, select appropriate sound acquisition equipment, such as a high-sensitivity microphone array, and install it at strategic locations around the circuit breaker. The sampling frequency of the acquisition equipment is set to 44.1kHz to ensure that it can capture all sound signals within the human audible range. The acquisition period is set to 5 seconds after each operation, which is sufficient to cover the entire circuit breaker operation process. During the acquisition process, a synchronous trigger device is used to ensure that sound signal acquisition begins synchronously with the circuit breaker operation. To obtain a comprehensive data set, multiple acquisitions are required under different operating conditions, including normal operation, slight wear, severe wear, and mechanical jams. At least 100 samples are collected for each condition to ensure statistical significance. The collected sound signals are saved in WAV format to ensure the integrity of the original data. A detailed database is also established to record metadata such as the environmental conditions, circuit breaker status, and acquisition time of each acquisition, providing important reference information for subsequent data analysis. The main purpose of this step is to obtain comprehensive and accurate original sound data, laying the foundation for subsequent feature extraction and fault diagnosis.

[0142] The specific implementation method of step S20 is: pre-process the collected original sound signal to improve the quality and reliability of the signal. First, use a bandpass filter to filter the signal, and the passband range of the filter is set to 100Hz to 10kHz. This range covers the frequency components of most mechanical fault sounds. The filter is designed using a Butterworth filter, and the order is selected as 8th order to ensure good roll-off characteristics. After filtering, the wavelet threshold denoising method is used for noise reduction. The db4 wavelet basis is selected, and the number of decomposition layers is 5. The wavelet coefficients of each layer are processed using a soft threshold function, and the threshold is selected using the Stein unbiased risk estimation criterion. The specific formula is: Where σ is the standard deviation of the noise and N is the signal length. During the noise reduction process, the mutation segment of the signal is detected at the same time. The mutation segment detection function is defined as Where s(t) is the original signal, is the local mean of the signal. When D(t) exceeds the preset threshold (recommended to be set to 3 times the standard deviation of the signal), the signal segment is marked as an abnormal mutation segment, and its start time, duration, and amplitude information are recorded. These abnormal mutation segments are excluded from the preprocessed signal and saved separately as an abnormal waveform segment set for subsequent feature fusion. The preprocessed signal and the abnormal waveform segment set are saved as independent data files for subsequent processing. The main purpose of this step is to improve the signal-to-noise ratio of the signal, remove irrelevant noise interference, and retain the abnormal mutation features that may contain important fault information.

[0143] The specific implementation method of step S30 is: in order to improve the efficiency of data processing, this step adopts the collaborative working mode of GPU and CPU. First, initialize the NVIDIA CUDA environment and set the GPU device parameters, including thread block size (recommended to be set to 256) and grid size (dynamically adjusted according to the amount of data). Create a CUDA stream object to manage parallel tasks on the GPU. Copy the preprocessed sound signal data from the host memory to the global memory of the GPU, and use the CUDAMemcpyAsync function to implement asynchronous data transmission to maximize GPU utilization. At the same time, create an input stream object on the CPU side to receive the abnormal waveform segment set. The CPU input stream adopts a double buffering mechanism and sets the buffer size to 1MB to balance memory usage and data reading efficiency. In order to ensure the integrity and consistency of the data, a synchronization mechanism is implemented, using conditional variables and mutex locks to coordinate the data processing progress of the GPU and CPU. On the GPU side, implement a CUDA kernel function for parallel processing of operations such as short-time Fourier transform of preprocessed signals, preferably the Fourier transform function of CUDA. On the CPU side, a multi-threaded processing function is implemented to process the feature extraction of abnormal waveform segment sets, preferably a multi-threaded Fourier transform function. The number of CPU threads is set to twice the number of system logical cores to make full use of CPU resources. Through this GPU-CPU collaborative processing method, the speed of data processing can be significantly improved, especially for large-scale data sets. Especially in actual use, the number of abnormal mutation segments filtered is small but the spectral characteristics are relatively complex, and the number of pre-processed sound signals after filtering is large but the spectral characteristics are relatively simple. Therefore, using the CPU to process abnormal mutation segments with a small amount but relatively complex spectral characteristics, and using the GPU to process pre-processed sound signals after filtering with a large amount but relatively simple spectral characteristics, can well distribute GPU-CPU collaborative processing and improve overall processing efficiency. The main purpose of this step is to prepare an efficient parallel computing environment for the subsequent feature extraction process and improve the processing efficiency of the entire system.

[0144] The specific implementation of step S40 is: using the parallel processing capability of the GPU to perform multiple feature extraction on the pre-processed sound signal. First, a short-time Fourier transform (STFT) is performed. The Hanning window is selected as the window function, the window length is set to 1024 sampling points, and the overlap rate is 50%. The calculation formula of STFT is: Where x[n] is the input signal, w[n] is the window function, m is the time index, and ω is the angular frequency. Next, the Mel-Frequency Cepstral Coefficients (MFCCs) are calculated. First, the STFT result is converted to the Mel-frequency scale using a 40-Mel filter bank. The Mel-frequency conversion formula is: Then, the logarithm of the filter output is taken and discrete cosine transform is performed, and the first 13 coefficients are retained as MFCC features. Finally, wavelet packet decomposition is performed. The db4 wavelet basis is selected and the number of decomposition layers is 5. The energy feature E is calculated for each subband. i =∑ k |d i [k]| 2 , where d i [k] is the wavelet coefficient of the i-th subband. The STFT, MFCC, and wavelet packet energy features are combined into a feature matrix, called the first feature matrix. Each row of the matrix represents a time frame, and the columns contain different types of features. To improve the distinguishability of the features, the first feature matrix is ​​normalized using the Z-score normalization method: Where μ is the feature mean and σ is the feature standard deviation. The main purpose of this step is to extract multi-dimensional and multi-scale sound features from the preprocessed signal to provide rich feature representation for subsequent fault identification.

[0145] The specific implementation of step S50 is: cluster analysis is performed on the first feature matrix to extract typical sound features. Clustering is performed using the improved K-means++ algorithm. First, the optimal number of clusters K is determined using the elbow rule. The overall intra-class sum of squares (WSS) under different K values ​​is calculated, and the inflection point where the WSS decline rate slows down significantly is selected as the optimal K value, which is generally between 5 and 10. When initializing the cluster center, the first center is randomly selected, and the selection probability of subsequent centers is proportional to the square of their minimum distance to the selected center. The iterative update formula of the clustering process is: where c i is the i-th cluster center, S i is the set of samples belonging to class i. To improve clustering stability, a strategy of averaging multiple runs is adopted, specifically 10 runs, and the result with the smallest sum of squares within the overall class is selected. To deal with possible outliers, in each iteration, the distance from each sample to its nearest center is calculated. Samples with a distance greater than 3 times the standard deviation are marked as potential outliers and given a smaller weight when updating the cluster center. After clustering is completed, the representative score of each cluster center is calculated. The scoring formula is: where σ i is the average distance of samples within the i-th cluster. The cluster centers with the highest scores are selected as representative sound features. The main purpose of this step is to extract the most representative sound feature patterns from a large amount of feature data, simplify the subsequent feature matching process, and improve the efficiency and accuracy of fault identification.

[0146] The specific implementation of step S60 is: based on the cluster centers obtained in step S50, the first feature matrix is ​​subjected to dimensionality reduction and feature replacement. First, the distance from each sample point to each cluster center is calculated, and the distance calculation adopts the Euclidean distance: Where x is the sample point, c i is the i-th cluster center, and n is the feature dimension. Then, for each sample point, find the closest cluster center and replace the coordinates of the original sample point with the coordinates of the cluster center. In order to retain certain local structural information, a soft allocation mechanism is introduced. Define the soft allocation weight: Where T is the temperature parameter, which controls the degree of soft assignment and is recommended to be set to the average value of the distance. The new coordinates of the sample point are calculated as: This soft allocation mechanism can retain the main features while also partially retaining the information of the secondary features. In order to further reduce the redundancy of the data, principal component analysis (PCA) is performed on the replaced feature matrix. The covariance matrix is ​​calculated: Where m is the number of samples, is the sample mean. Solve the characteristic equation Cv = γv to obtain the eigenvalues ​​and eigenvectors. Select the first k principal components with a cumulative contribution rate of 95%, project the data into the subspace formed by these k principal components, and obtain the second characteristic matrix. Finally, normalize the second characteristic matrix using the Min-Max normalization method: The main purpose of this step is to significantly reduce the amount of data and improve the efficiency of subsequent processing while retaining key information through feature replacement and dimensionality reduction.

[0147] The specific implementation method of step S70 is: process the set of abnormal waveform segments and align their features with the second feature matrix. First, feature extraction is performed on each waveform segment in the abnormal waveform segment set. The same method as step S40 is used to calculate the STFT, MFCC and wavelet packet energy features. Taking into account the short-term characteristics of the abnormal waveform segments, the calculation of STFT and MFCC is adjusted accordingly: the window length is reduced to 512 sampling points, and the overlap rate is increased to 75%. For wavelet packet decomposition, the 5-layer decomposition is maintained, but only the energy features of the high-frequency sub-band (4th and 5th layers) are selected to highlight the abnormal features. The extracted features are combined into a third feature matrix. In order to align with the second feature matrix, alignment of the time and feature dimensions is required. Time alignment: Use the timestamp information of the abnormal waveform segment to map it to the time frame of the second feature matrix. If an abnormal waveform segment spans multiple time frames, its features are distributed to the corresponding frames according to the time ratio. Feature dimension alignment: Since the dimension of the third feature matrix may be different from that of the second feature matrix, dimensional adjustment is required. The third characteristic matrix is ​​upsampled or downsampled using linear interpolation to make its dimension consistent with the second characteristic matrix. The interpolation formula is: Among them, (x1, y1) and (x2, y2) are adjacent known points, (x i ,y i ) is the point to be interpolated. After alignment, the third feature matrix is ​​normalized using the same Min-Max normalization method as the second feature matrix. In order to enhance the expression of abnormal features, the attention mechanism is introduced. Calculate the abnormality score: where x i is the i-th eigenvector, is the average feature vector, σ is the standard deviation. Construct attention weights based on the scores: The attention weights are applied to the third feature matrix to obtain the weighted feature matrix. The main purpose of this step is to effectively integrate the feature information of the abnormal waveform segment with the features of the normal signal, providing a more comprehensive information basis for subsequent feature fusion.

[0148] The specific implementation of step S80 is: the second feature matrix and the third feature matrix are fused and dimensionality reduction is performed. First, the feature fusion is performed using a preset feature fusion equation group. The time-frequency domain fusion adopts the equation: The initial values ​​of α, β, γ, and δ are set to 0.4, 0.3, 0.2, and 0.1, and then optimized using the particle swarm optimization algorithm. The parameter settings for particle swarm optimization are: population size is 50, maximum number of iterations is 100, inertia weight w is linearly decreased from 0.9 to 0.4, and acceleration constant c1 = c2 = 2. The optimization objective function is the ratio of the inter-class distance to the intra-class distance of the fused features. Energy distribution fusion uses the equation: Among them, w i The entropy weight method is used to calculate η, which is initially set to 0.2 and subsequently optimized through cross-validation. The phase spectrum fusion uses the equation: Among them, v i Obtained through phase consistency evaluation, μ is initially set to 0.1 and optimized through grid search. Abnormal feature fusion uses the equation: After fusion, a high-dimensional feature matrix is ​​obtained. Next, principal component analysis (PCA) is used for dimensionality reduction. First, the fused feature matrix is ​​centered: in is the mean of each feature. Calculate the covariance matrix: Where n is the number of samples. Solve the characteristic equation Cv=λv and get the characteristic value λ i and the corresponding eigenvector v i Arrange the eigenvalues ​​in descending order and select the first k principal components whose cumulative contribution rate reaches 95%. Contribution rate calculation formula: Construct the projection matrix P, whose columns are the selected eigenvectors. Finally, through X reduced =X centered P obtains the feature matrix after dimensionality reduction. In order to improve the robustness of PCA, the incremental PCA algorithm is considered to process large-scale data sets in batches. The update formula of incremental PCA is: Where t is the current sample number, x t The main purpose of this step is to make full use of information from different sources through feature fusion, remove redundant information through dimensionality reduction, and extract the most representative feature combination.

[0149] The specific implementation of step S90 is: standardize the feature matrix after dimensionality reduction to obtain the final mechanical fault voiceprint feature matrix. First, check whether there are outliers in the feature matrix after dimensionality reduction. Use the box plot method to identify outliers, and regard values ​​less than Q1-1.5IQR or greater than Q3+1.5IQR as potential outliers, where Q1 and Q3 are the first and third quartiles respectively, and IQR is the interquartile range. For the detected outliers, instead of simply deleting them, a winsorization method is used to limit the outliers to the range of [Q1-1.5IQR, Q3+1.5IQR]. Then, the processed feature matrix is ​​Z-score standardized: Where μ is the feature mean and σ is the feature standard deviation. To improve the robustness of standardization, consider using the median and median absolute deviation (MAD) instead of the mean and standard deviation: Where MAD(x) = median(|x - median(x)|). After standardization, the correlation between features is examined, and the correlation matrix is ​​calculated using the Pearson correlation coefficient. If there are highly correlated feature pairs (with an absolute value of the correlation coefficient greater than 0.9), consider removing one of the features to reduce multicollinearity. To enhance feature interpretability, the standardized features are named and described. For example, features can be named based on their source (time-frequency domain, energy distribution, phase spectrum, abnormal characteristics) and physical meaning (frequency center, energy concentration, phase consistency, etc.). Finally, a feature importance ranking is constructed, which can be based on the feature's variance contribution or using feature importance scores from models such as random forests. This feature importance information is appended to the final feature matrix so that subsequent fault diagnosis models can more specifically utilize these features. The main purpose of this step is to standardize the reduced features, enabling direct comparison of features of different dimensions. Furthermore, through outlier processing and feature selection, the quality and reliability of the features are further improved, providing high-quality input data for subsequent fault diagnosis and classification tasks.

[0150] Through steps S10 to S90, as described in detail above, the entire process from raw sound signal acquisition to the final mechanical fault soundprint feature matrix is ​​completed. This process encompasses key steps such as signal acquisition, preprocessing, feature extraction, clustering analysis, feature fusion, dimensionality reduction, and normalization. Each step is carefully designed to maximize the extraction and preservation of acoustic feature information related to circuit breaker mechanical faults. By utilizing a variety of advanced signal processing and machine learning techniques, such as wavelet analysis, deep learning, clustering algorithms, and principal component analysis, the most representative and discriminative features are extracted from complex sound signals. The resulting mechanical fault soundprint feature matrix not only contains rich time-frequency domain information, energy distribution characteristics, phase characteristics, and anomaly features, but also undergoes dimensionality reduction and normalization, making it more suitable for subsequent fault diagnosis and classification tasks. The innovation of this method lies in its comprehensive consideration of acoustic features under normal operation and various fault conditions, effectively integrating this information through feature fusion. By incorporating GPU parallel processing and CPU collaboration, this method also achieves high computational efficiency and is capable of processing large amounts of sound data. This systematic and comprehensive feature extraction method provides strong technical support for mechanical fault diagnosis of circuit breakers, and is expected to significantly improve the accuracy and timeliness of fault detection, making an important contribution to the safe and stable operation of power systems.

[0151] Specifically, the principle of the present invention is:

[0152] 1. Principle of multimodal feature fusion: Mechanical failures of circuit breakers often manifest as complex sound patterns, and a single type of feature is difficult to fully capture fault information. The present invention constructs a multi-dimensional feature space by fusing time-frequency domain features, energy distribution features, phase spectrum features, and abnormal features. The time-frequency domain features reflect the time-varying frequency characteristics of the signal, the energy distribution features characterize the energy proportion of different frequency bands, the phase spectrum features capture the phase change law of the signal, and the abnormal features focus on describing transient fault information. This multimodal fusion strategy ensures the comprehensiveness of feature extraction and can more accurately describe various complex fault modes.

[0153] 2. Adaptive Signal Preprocessing Principle: Considering that circuit breaker sound signals often contain complex background noise, this invention employs an adaptive preprocessing method. A bandpass filter is used to initially remove irrelevant frequency bands, while wavelet threshold noise reduction adaptively adjusts the threshold based on the local characteristics of the signal, effectively removing noise while preserving signal details. In particular, the separate processing of abnormal and sudden changes demonstrates a special focus on transient fault information. This differentiated processing strategy ensures that critical fault information is not lost during the preprocessing process.

[0154] 3. Parallel Computing Acceleration Principle: Based on the characteristics of modern computing hardware, this invention adopts a GPU-CPU co-processing architecture. The GPU's massively parallel processing capabilities are particularly well-suited for computationally intensive tasks such as Fourier transforms and wavelet decomposition, while the CPU is more suitable for tasks requiring complex control flows, such as processing abnormal waveform segments. This co-processing mechanism fully leverages the advantages of heterogeneous computing platforms, significantly improving the efficiency of feature extraction.

[0155] 4. Feature Space Optimization Principle: To improve feature expressiveness and reduce redundancy, this invention incorporates cluster analysis and principal component analysis. Cluster analysis identifies similar feature patterns to extract the most representative sound features, reducing data redundancy. The introduction of a soft allocation mechanism preserves primary features while also partially retaining information about secondary features, improving feature robustness. Principal component analysis further optimizes the feature space by finding the direction of maximum variance, removing redundant information and improving feature discrimination.

[0156] 5. Principle of nonlinear feature fusion: The feature fusion equation group proposed in this invention adopts a nonlinear fusion strategy. The time-frequency domain fusion equation introduces the second-order partial derivative term to enhance the capture of edge information. The energy distribution fusion equation adopts the form of entropy to better describe the degree of energy dispersion. The phase spectrum fusion equation takes into account the cyclic characteristics of the phase and avoids the problems caused by phase jumps. The abnormal feature fusion equation adopts the form of Gaussian function, which can more accurately describe the time distribution of mutation features. This nonlinear fusion strategy can better capture the complex relationship between features and improve the expressive power of features.

[0157] 6. Adaptive Parameter Optimization Principle: This invention incorporates parameter optimization mechanisms at multiple stages. For example, during the feature fusion process, a particle swarm optimization algorithm is used to dynamically adjust fusion weights to accommodate different fault signatures. In cluster analysis, the elbow rule is used to adaptively determine the optimal number of clusters. This adaptive parameter optimization strategy makes the method more versatile and adaptable, enabling it to handle circuit breaker failures under diverse operating conditions.

[0158] In summary, the technical solution of the present invention constructs a comprehensive, efficient and accurate circuit breaker mechanical fault voiceprint feature extraction framework through the organic combination of multimodal feature fusion, adaptive signal preprocessing, parallel computing acceleration, feature space optimization, nonlinear feature fusion and adaptive parameter optimization.

[0159] The following is an example of a specific application scenario of the present invention: A provincial power company implemented the circuit breaker mechanical fault voiceprint feature extraction system based on the present invention at its main substation. The substation had 50 10kV circuit breakers, model ZN63A-12, ranging in age from 2 to 15 years. To verify the effectiveness of the present method, the power company selected 10 circuit breakers of varying ages for a six-month trial run.

[0160] 1. Data collection stage

[0161] First, staff installed four high-sensitivity microphones (model: PCB378B02) around each selected circuit breaker, with a sampling frequency set to 44.1kHz. The microphones were positioned in front of, behind, to the left, and to the right of the circuit breaker, approximately 0.5 meters from the breaker itself. A synchronous triggering system was also installed to ensure that sound signal acquisition began synchronously with circuit breaker operation.

[0162] During the six-month trial operation, the system recorded the normal operating sound of each circuit breaker and various possible fault sounds. The specific collection details are as follows:

[0163] -Normal operation sound: Each circuit breaker is routinely operated at least five times a week, and a total of approximately 1,300 normal operation sounds are collected.

[0164] - Mild wear state: The contact gap is manually adjusted to simulate a mild wear state, and data is collected 50 times for each circuit breaker.

[0165] -Severe wear state: Severe wear state is simulated by replacing some worn parts, and each circuit breaker is collected 30 times.

[0166] -Mechanical stuck state: The mechanical stuck state was simulated by artificially increasing the friction of the operating mechanism. Each circuit breaker was collected 20 times.

[0167] All collected sound signals are saved in WAV format, and detailed metadata is recorded, including circuit breaker number, operation time, ambient temperature, humidity and other information.

[0168] 2. Data preprocessing stage

[0169] The collected raw sound signal is preprocessed. First, a bandpass filter is used with a passband range of 100 Hz to 10 kHz. An 8th-order Butterworth filter is used. Noise reduction is then performed using a five-layer wavelet threshold denoising method based on the DB4 wavelet.

[0170] During the noise reduction process, the system detected multiple abnormal mutation segments. Taking one of the circuit breaker (No.: CB-001) as an example, three significant abnormal mutation segments were detected:

[0171] 1. Time: 0.15s-0.18s, Amplitude: 1.2 times the standard deviation

[0172] 2. Duration: 0.42s-0.46s, Amplitude: 2.8 times the standard deviation

[0173] 3. Time: 0.89s-0.91s, Amplitude: 1.5 times the standard deviation

[0174] These abnormal mutation segments are saved separately as abnormal waveform segment sets for subsequent feature fusion.

[0175] 3. Parallel processing configuration

[0176] The system was configured on a high-performance server equipped with NVIDIA Tesla V100 GPUs and dual-socket Intel Xeon Gold 6248R CPUs. During CUDA environment initialization, the thread block size was set to 256, and the grid size was dynamically adjusted based on the data volume. Forty-eight threads (corresponding to 24 physical cores) were created on the CPU to process the set of abnormal waveform segments.

[0177] 4. Feature extraction stage

[0178] Multi-dimensional feature extraction is performed on the preprocessed signal. Taking a normal operation of the CB-001 circuit breaker as an example, the extracted features are as follows:

[0179] Short-time Fourier transform (STFT):

[0180] -Window length: 1024 samples

[0181] -Overlap rate: 50%

[0182] -Frequency resolution: approximately 43Hz

[0183] -Time resolution: about 11.6ms

[0184] Mel-frequency cepstral coefficients (MFCC):

[0185] - Number of Mel filters: 40

[0186] -Retention coefficient number: 13

[0187] Wavelet packet decomposition:

[0188] - Wavelet basis: db4

[0189] -Number of decomposition layers: 5

[0190] -Number of extracted sub-band energy features: 32

[0191] 5. Cluster Analysis

[0192] Cluster analysis was performed on the extracted features. Using the improved K-means++ algorithm, the optimal number of clusters was determined using the elbow rule. After multiple experiments, the optimal number of clusters was determined to be K = 7. The clustering results are shown in Table 1:

[0193] Table 1 Clustering results

[0194] Cluster number Sample size Main feature description 1 3245 Low-frequency energy is concentrated and MFCC coefficients are stable 2 2876 The intermediate frequency energy is prominent and the phase spectrum fluctuates greatly 3 1987 Rich high-frequency components and uniform wavelet energy distribution 4 1532 The time-frequency characteristics change dramatically, which may be related to mechanical jamming. 5 1245 The MFCC coefficient changes significantly, which may indicate a contact problem. 6 987 The low-frequency energy decays slowly, and there may be a problem with the operating mechanism. 7 654 High frequency bursts, possibly related to arc faults

[0195] 6. Feature replacement and dimensionality reduction

[0196] Based on the clustering results, feature replacement was performed on each sample point. A soft assignment mechanism was used, with the temperature parameter T set to the average distance (approximately 0.35). After feature replacement, PCA dimensionality reduction was performed on the data. Principal components with a cumulative contribution rate of 95% were selected, ultimately reducing the feature dimension from the original 85 to 23 dimensions.

[0197] 7. Feature Fusion

[0198] The preset feature fusion equations are applied to perform feature fusion. Taking time-frequency domain fusion as an example, the fusion weight coefficients are obtained by particle swarm optimization algorithm: α = 0.42, β = 0.28, γ = 0.18, δ = 0.12.

[0199] In energy distribution fusion, the weight coefficient η of the energy of the abnormal mutation segment is set to 0.2. In phase spectrum fusion, the weight coefficient μ of the phase of the abnormal mutation segment is set to 0.15.

[0200] 8. Final feature matrix construction

[0201] After normalization and correlation analysis, an 18-dimensional mechanical fault voiceprint feature matrix is ​​finally obtained. The features of each dimension and their importance scores are shown in Table 2:

[0202] Table 2 Dimensional feature table

[0203]

[0204]

[0205] 9. System performance evaluation

[0206] To evaluate the performance of our method, we compared the fault detection accuracy and computation time between a traditional method (using only time-domain and frequency-domain features) and our method. The test dataset contained 2,000 normal samples and 500 samples of various fault types. The results are shown in Table 3.

[0207] Table 3 Performance comparison table

[0208] Performance indicators Traditional methods Method of the present invention Improvement Fault detection accuracy 87.5％ 96.8％ 10.6％ Average computation time (per sample) 2.3s 0.8s 65.2％ False positive rate 8.2％ 2.5％ 69.5％ False negative rate 6.8％ 1.8％ 73.5％

[0209] 10. Typical failure case analysis

[0210] During the trial run, the system successfully detected several potential faults. The following is a typical case:

[0211] Circuit breaker number: CB-007 (service life: 8 years)

[0212] Detection time: July 15, 2024, 14:32:45

[0213] Environmental conditions: temperature 28°C, humidity 65%

[0214] The system detected an abnormality in the sound characteristics of the circuit breaker. The main abnormalities are manifested in the following characteristic dimensions:

[0215] 1. The proportion of mid-frequency energy (feature 2) increased significantly, from a normal level of 0.28 to 0.41.

[0216] 2. The fluctuation of the second coefficient of MFCC (feature 6) increases, and the standard deviation increases from 0.03 to 0.09.

[0217] 3. The energy of the high-frequency subband of the wavelet packet (feature 9) suddenly increases, with an amplitude exceeding 2.5 times the normal level.

[0218] 4. Both the intensity (feature 13) and duration (feature 14) of abnormal mutations increased significantly.

[0219] Based on these anomalies, the system inferred that the circuit breaker might have excessive contact wear. The power company immediately arranged for an inspection, which confirmed the severe wear and replaced the relevant components promptly, preventing a potentially more serious failure.

[0220] The above description is only a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any technician familiar with this technical field can easily think of changes or replacements within the technical scope disclosed by the present invention, which should be covered by the scope of protection of the present invention.

Claims

1. A method for extracting voiceprint features of mechanical faults of circuit breakers, characterized in that: The following steps are involved: S10, continuously collect sound signals during normal operation of the circuit breaker and various mechanical fault conditions; S20, filtering and denoising the sound signal to obtain a preprocessed sound signal and abnormal sudden change segments of the sound signal waveform excluded during the filtering and denoising process, which are recorded as an abnormal waveform segment set; S30, starting the CUDA parallel processing flow of the parallel computing platform of the graphics processing unit GPU to receive the pre-processed sound signal; Starting an input stream of a central processing unit (CPU) to receive the abnormal waveform segment set; S40, using a sliding window to perform short-time Fourier transform, Mel-frequency cepstral coefficient analysis, and wavelet packet decomposition on the preprocessed sound signal to obtain features of the preprocessed sound signal, which are recorded as a first feature matrix; S50, clustering the first feature matrix to obtain multiple cluster centers for representing typical sound features; S60, for each element in the cluster, replacing the element with the cluster center to obtain a second feature matrix; S70, extracting features of the abnormal waveform segment set from the second feature matrix to obtain a third feature matrix, and aligning the second feature matrix and the third feature matrix based on the second feature matrix; S80, fusing the second feature matrix and the third feature matrix using a preset feature fusion equation group, including a time-frequency domain fusion equation, an energy distribution fusion equation, a phase spectrum fusion equation, and an abnormal feature fusion equation, and performing dimensionality reduction using principal component analysis to obtain a feature matrix after dimensionality reduction; S90, performing normalization processing on the feature matrix after dimensionality reduction, and finally obtaining a mechanical fault voiceprint feature matrix; The time-frequency domain fusion equation is specifically expressed as: Among them: F t (x, y) is the fused time-frequency domain feature; S(x, y) is the time-frequency domain feature in the second feature matrix; T(x, y) is the time-frequency domain feature in the third feature matrix; α, β, γ, δ are weight coefficients obtained by particle swarm optimization algorithm; is the second-order partial derivative of S(x,y), which is used to enhance edge information; A(x,y) is the time-frequency feature of the abnormal mutation segment, obtained by short-time Fourier transform; The energy distribution fusion equation is specifically expressed as follows: Where: E(f) is the energy distribution characteristic after fusion; P i (f) is the power spectral density of the i-th feature matrix; w i is the weight coefficient, calculated by entropy weight method; N is the number of feature matrices; η is the weight coefficient of abnormal mutation segment energy; M is the number of abnormal mutation segments; D j is the duration of the jth abnormal mutation segment; T j is the total duration of the signal; E j (f) is the energy spectrum density of the jth abnormal mutation segment; The phase spectrum fusion equation is specifically expressed as: Where: Φ(f) is the phase spectrum feature after fusion; φ i (f) is the phase spectrum of the i-th characteristic matrix; v i is the weight coefficient, obtained through phase consistency evaluation; k is an integer used to ensure phase continuity; μ is the weight coefficient of the abnormal mutation phase; A j is the mutation amplitude of the jth abnormal mutation segment; A max is the maximum mutation amplitude among all abnormal mutation segments; φ j (f) is the phase spectrum of the jth abnormal mutation segment; The abnormal feature fusion equation is specifically expressed as: Among them: F a (t) is the abnormal feature after fusion; M is the number of abnormal mutation segments; λ j is the weight coefficient of the jth abnormal mutation segment, which is calculated comprehensively by the mutation amplitude and duration; t j,start and t j,end are the start and end time of the jth abnormal mutation segment; t j,mid is the midpoint of the jth abnormal mutation segment; σ j is the time scale parameter of the jth abnormal mutation segment, which is related to the duration; A j is the mutation amplitude of the jth abnormal mutation segment.

2. A circuit breaker mechanical fault soundprint feature extraction method according to claim 1, characterized in that: The clustering is achieved using an improved K-means++ algorithm.

3. The method for extracting circuit breaker mechanical fault soundprint features according to claim 2, characterized in that: The time-frequency characteristics of the abnormal mutation segment are specifically calculated as follows: Among them, s a (t) is the abnormal mutation segment signal, w(t) is the window function, and t represents time.

4. The method for extracting circuit breaker mechanical fault soundprint features according to claim 3, characterized in that: The power spectrum density of the characteristic matrix is ​​specifically calculated as follows: Among them, s i (t) is the time domain signal corresponding to the i-th feature matrix, and T is the signal duration.

5. The method for extracting circuit breaker mechanical fault soundprint features according to claim 4, characterized in that: The energy distribution characteristics after fusion are specifically calculated as follows: Among them, s j (t) is the signal of the jth abnormal mutation segment.

6. A circuit breaker mechanical fault soundprint feature extraction method according to claim 5, characterized in that: The fused phase spectrum feature is specifically calculated as follows: Among them, X i (f) is the Fourier transform of the i-th eigenvalue matrix.

Images