Transformer partial discharge identification method and electronic equipment

Abstract

The invention discloses a partial discharge identification method of a transformer and electronic equipment. Relates to the field of intelligent power grids, and the method comprises the following steps: obtaining partial discharge sound signals corresponding to a plurality of channels in a microphone array, the partial discharge sound signals being acquired from a transformer; based on the partial discharge sound signals corresponding to the multiple channels, target spectrograms corresponding to the multiple channels are obtained, and the target spectrograms represent energy distribution conditions of the corresponding partial discharge sound signals at multiple frequencies and multiple times; obtaining a target partial discharge sound signal based on the target spectrograms corresponding to the plurality of channels; and determining a target partial discharge type of the transformer based on the target partial discharge sound signal. According to the invention, the technical problem of insufficient transformer partial discharge identification accuracy caused by incomplete consideration of factors in a complex field environment is solved.

Description

Technical Field

[0001] This invention relates to the field of smart grids, and more specifically, to a method and electronic device for identifying partial discharge in transformers. Background Technology

[0002] Transformers are critical equipment in power systems, and their insulation condition directly affects the safe and stable operation of the power grid. Partial discharge, as a common early insulation fault phenomenon, generates acoustic signals that are important clues for monitoring the health status of transformers. In recent years, with the increasing demand for intelligent power equipment, real-time monitoring and accurate identification of partial discharge activity have become particularly important. The significant shortcomings of related technologies in transformer partial discharge identification are mainly reflected in the following aspects:

[0003] Single-channel spectrum analysis methods struggle to utilize redundant information from multi-channel data to suppress noise and find it difficult to balance time-frequency resolution. Fixed-threshold detection methods rely excessively on preset thresholds and are ill-suited to the time-varying nature of signals. Furthermore, convolutional neural networks or residual network methods employed in related technologies, due to their fixed kernel size and high computational complexity, are ill-suited to meet the requirements of real-time performance and embedded deployment.

[0004] Although relevant partial discharge audio signal recognition technologies can provide basic identification and monitoring of partial discharge within a certain range, when faced with multiple types of partial discharge signals in complex field environments, the inability to effectively integrate multi-channel information and cross-domain features leads to insufficient accuracy in identifying transformer partial discharge.

[0005] There is currently no effective solution to the above problems. Summary of the Invention

[0006] This invention provides a method and electronic device for identifying partial discharge in transformers, which at least solves the technical problem of insufficient accuracy in identifying partial discharge in transformers due to incomplete consideration of factors when facing complex field environments.

[0007] According to one aspect of the present invention, a method for identifying partial discharge in a transformer is provided, comprising: acquiring partial discharge sound signals corresponding to multiple channels in a microphone array, wherein the partial discharge sound signals are acquired from the transformer; obtaining target spectra corresponding to multiple channels based on the partial discharge sound signals corresponding to the multiple channels, wherein the target spectra represent the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times; obtaining target partial discharge sound signals based on the target spectra corresponding to the multiple channels; and determining the target partial discharge type of the transformer based on the target partial discharge sound signals.

[0008] According to another aspect of the present invention, a partial discharge identification device for a transformer is also provided, comprising: a partial discharge sound signal acquisition module, configured to acquire partial discharge sound signals corresponding to each of multiple channels in a microphone array, wherein the partial discharge sound signals are acquired from the transformer; a target spectrum determination module, configured to obtain target spectra corresponding to each of the multiple channels based on the partial discharge sound signals corresponding to each of the multiple channels, wherein the target spectrum represents the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times; a target partial discharge sound signal determination module, configured to obtain a target partial discharge sound signal based on the target spectrum corresponding to each of the multiple channels; and a target partial discharge type determination module, configured to determine the target partial discharge type of the transformer based on the target partial discharge sound signal.

[0009] According to another aspect of the present invention, a non-volatile storage medium is also provided, the non-volatile storage medium storing a plurality of instructions adapted for loading by a processor and executing any one of the transformer partial discharge identification methods described herein.

[0010] According to another aspect of the present invention, an electronic device is also provided, including one or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the partial discharge identification method for transformers as described in any one of the present invention.

[0011] According to another aspect of the present invention, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the partial discharge identification method for a transformer as described in any one of the present invention.

[0012] In this embodiment of the invention, partial discharge sound signals corresponding to multiple channels in a microphone array are acquired, wherein the partial discharge sound signals are collected from the transformer; based on the partial discharge sound signals corresponding to each of the multiple channels, target spectra corresponding to each of the multiple channels are obtained, wherein the target spectra represent the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times; based on the target spectra corresponding to each of the multiple channels, target partial discharge sound signals are obtained; based on the target partial discharge sound signals, the target partial discharge type of the transformer is determined. This achieves the goal of obtaining the target spectra by processing the acquired multi-channel partial discharge sound signals, thereby obtaining the target partial discharge sound signals and accurately determining the target partial discharge type of the transformer. This improves the accuracy of transformer partial discharge identification and solves the technical problem of insufficient accuracy of transformer partial discharge identification caused by incomplete consideration of factors in complex field environments. Attached Figure Description

[0013] The accompanying drawings, which are included to provide a further understanding of the invention and form part of this application, illustrate exemplary embodiments of the invention and, together with their description, serve to explain the invention and do not constitute an undue limitation thereof. In the drawings:

[0014] Figure 1 This is a flowchart of a partial discharge identification method for a transformer according to an embodiment of the present invention;

[0015] Figure 2 This is a flowchart of an optional transformer partial discharge identification method according to an embodiment of the present invention;

[0016] Figure 3 This is a schematic diagram of a partial discharge identification device for a transformer according to an embodiment of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] Partial discharge refers to the phenomenon in high-voltage power systems or electrical equipment where, when the voltage in the insulating medium reaches a certain level, the discharge occurs only in a localized area of ​​the insulating medium, rather than throughout the entire gap. This discharge activity is usually caused by the presence of tiny voids, cracks, or other inhomogeneities in the insulating material, and it can occur in solid, liquid, or gaseous insulating media.

[0020] Spectral masking is a signal processing technique primarily used to extract target signals from noisy signals. Spectral masks are widely used in tasks such as noise reduction, enhancing signal clarity, and signal recognition.

[0021] According to an embodiment of the present invention, a method for identifying partial discharge of a transformer is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions. Furthermore, although a logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order than that shown here.

[0022] Figure 1 This is a flowchart of a partial discharge identification method for a transformer according to an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes the following steps:

[0023] Step S102: Acquire the partial discharge sound signals corresponding to each of the multiple channels in the microphone array, wherein the partial discharge sound signals are acquired from the transformer.

[0024] Optionally, the microphone array consists of multiple microphones distributed in a specific geometry in physical space to simultaneously capture signals from the same source from different angles or positions. In transformer partial discharge detection, each microphone (channel) in the array independently records the acoustic signals of the environment surrounding the transformer. These signals include the sound waves of partial discharge and other environmental noise. When a partial discharge event occurs, energy is released in the form of sound waves, which are captured by the microphone array. The fusion of multi-channel data can enhance the signal-to-noise ratio, effectively suppress background noise, and improve the accuracy and reliability of subsequent signal identification.

[0025] Step S104: Based on the partial discharge sound signals corresponding to each of the multiple channels, obtain the target spectrum corresponding to each of the multiple channels, wherein the target spectrum represents the energy distribution of the corresponding partial discharge sound signal at multiple frequencies and multiple times.

[0026] Optionally, the partial discharge sound signals acquired by each microphone channel can be converted into time-frequency spectra using short-time Fourier transform. Each point on these spectra represents the energy level of the partial discharge sound signal at a specific time and frequency. The goal is to capture more detailed information about the signal from the time domain to the time-frequency domain through this transformation, including the transient characteristics of the partial discharge pulse and its frequency components, thereby significantly improving the recognition accuracy of the partial discharge sound signal.

[0027] In one optional embodiment, obtaining target spectra for each of the multiple channels based on the partial discharge sound signals corresponding to each channel includes: performing bandpass filtering on the partial discharge sound signal of any channel to obtain a filtered partial discharge sound signal corresponding to any channel; normalizing the filtered partial discharge sound signal to obtain a normalized partial discharge sound signal corresponding to any channel; obtaining the target spectra for any channel based on the normalized partial discharge sound signal; and obtaining the target spectra for each of the multiple channels based on the partial discharge sound signals corresponding to each channel using the method of obtaining the target spectra for any channel.

[0028] Optionally, to separate and analyze these partial discharge sound signals from complex background noise, a series of signal processing steps are required. First, the raw partial discharge sound signals acquired by each microphone channel are bandpass filtered to effectively suppress low-frequency mechanical noise and high-frequency interference noise unrelated to partial discharge. The bandpass filtering of the partial discharge sound signal of any channel can be performed as follows to obtain the filtered partial discharge sound signal for any channel: ,in, This represents the filtered partial discharge sound signal corresponding to any channel. This represents the partial discharge sound signal of any channel, where i represents the index of any channel; This represents the impulse response of the bandpass filter, where k represents the order of the bandpass filter. , This indicates the preset minimum cutoff frequency of the passband. This indicates the preset minimum cutoff frequency of the passband. Next, to standardize the signal amplitude range and eliminate differences in signal strength between different channels, the filtered audio signal undergoes amplitude normalization processing to ensure that the signals from all channels are compared and analyzed on the same scale. The normalized partial discharge audio signal for any channel can be obtained by normalizing the filtered partial discharge audio signal in the following way: ,in, This represents the normalized partial discharge sound signal. This represents the mean value of the filtered partial discharge sound signal corresponding to any channel. This represents the standard deviation of the filtered partial discharge sound signal corresponding to any channel. Finally, short-time Fourier transform is used to perform time-frequency analysis on the normalized partial discharge sound signal to construct the target spectrum for each channel. The multi-channel target spectrum not only provides rich partial discharge features but also utilizes the spatial redundancy between multiple microphones to enhance the signal-to-noise ratio, further improving the ability to identify partial discharge signals in noisy environments.

[0029] In one optional embodiment, when there are multiple target spectra for any channel, obtaining the target spectra for any channel based on the normalized partial discharge sound signal includes: performing frame segmentation processing on the normalized partial discharge sound signal to obtain multiple frame signals for any channel, wherein the frame signal represents a signal segment of a predetermined sampling duration obtained from the normalized partial discharge sound signal; determining the target spectra corresponding to each of the multiple frame signals; and obtaining multiple target spectra for any channel based on the target spectra corresponding to each of the multiple frame signals for any channel.

[0030] Optionally, firstly, based on the normalized partial discharge sound signal, it is processed by framing, with each frame containing a fixed sampling duration. Framing allows for the capture of instantaneous changes in the signal. Next, a short-time Fourier transform (SFT) is performed on each frame, a time-frequency analysis technique that reveals the energy distribution of the signal at different frequencies within a specific time period. The SFT generates a target spectrum for each frame, providing the signal's time-frequency structure, including pulse characteristics and frequency components, which is crucial for identifying and classifying partial discharge types. Finally, by comprehensively considering the target spectra of all frames, a complete view of the partial discharge sound signal in the time-frequency domain is obtained, yielding richer and more detailed signal characteristics.

[0031] In one optional embodiment, the normalized partial discharge sound signal is subjected to frame segmentation processing to obtain multiple frame signals for any channel, including: performing frame segmentation processing on the normalized partial discharge sound signal to obtain any frame signal for any channel in the following manner:

[0032] ;

[0033] in, This represents any frame signal from any channel; i represents the index of any channel. The index represents the index of any frame signal, and L represents the number of frame signals; This represents the index of a sampling point in any frame of the signal. Indicates the scheduled sampling duration; This represents the normalized partial discharge sound signal; H represents the frame shift between the starting sampling points of two adjacent frame signals; multiple frame signals of any channel are obtained by obtaining any frame signal of any channel.

[0034] Optionally, framing is the process of dividing a continuous signal stream into a series of short-time signal segments. Framing divides the normalized partial discharge sound signal into a series of standardized short-time signal segments, laying a solid foundation for subsequent feature extraction, analysis, and recognition. This process not only improves processing efficiency but also enhances the ability to capture non-stationary signal features, especially in signal analysis against complex noise backgrounds, thereby improving the accuracy of partial discharge identification. It is a key step in achieving efficient and intelligent power equipment monitoring.

[0035] In one optional embodiment, determining the target spectrogram corresponding to each of the multiple frame signals includes: determining the target spectrogram of any frame signal in the following manner:

[0036] ;

[0037] in, This represents the target spectrum of any frame of signal; f represents the frequency. ; Indicates the index of any frame of signal. This represents the short-time Fourier transform result of any frame of signal; The Hanning window function is used to obtain the target spectrum of multiple frames of signals for any channel by using the method of obtaining the target spectrum of any frame signal.

[0038] Optionally, the time-domain signal can be converted to a time-frequency domain representation using a short-time Fourier transform. The target spectrum of each frame can provide the energy intensity of the signal at different frequencies, which is crucial for identifying the pulse characteristics and frequency distribution of partial discharges. These spectra preserve the time-frequency characteristics of the signal's transient changes, while reducing spectral leakage and improving signal purity through the application of the Hanning window function. The calculation of the target spectrum helps to accurately identify and distinguish partial discharge pulses in complex noisy environments, providing a strong foundation for subsequent feature extraction, classification, and decision-making.

[0039] Step S106: Based on the target spectrum corresponding to each of the multiple channels, obtain the target partial discharge sound signal.

[0040] Optionally, based on the target spectrum corresponding to each of the multiple channels, the obtained target partial discharge sound signal is not only highly clear, but also contains richer information on the type of partial discharge, providing a more accurate data foundation for the condition monitoring and fault diagnosis of power equipment.

[0041] In one optional embodiment, there are multiple target spectra corresponding to each channel, with multiple frame signals and multiple target spectra. Figure 1In the case of one-to-one correspondence, the target partial discharge sound signal is obtained based on the target spectra corresponding to each of the multiple channels, including: determining the fused spectra corresponding to each of the multiple frame signals based on the multiple target spectra corresponding to each of the multiple channels, wherein the fused spectra are obtained by fusing the target spectra corresponding to the same frame signal in the multiple target spectra corresponding to each of the multiple channels; determining the fused spectra mask corresponding to each of the multiple frame signals; and obtaining the target partial discharge sound signal based on the fused spectra corresponding to each of the multiple frame signals and the fused spectra mask corresponding to each of the multiple frame signals.

[0042] Optionally, firstly, based on multiple target spectra from multiple channels, a fused spectra for each frame of signal is determined. Since multi-channel fusion considers the signal differences that different sensors may capture, the resulting fused spectra better reflects the essential characteristics of the signal. Next, based on the signal energy distribution and characteristics in the fused spectra, specific algorithms, including but not limited to energy thresholding and pulse structure identification, are used to determine which time-frequency regions are more likely to contain partial discharge signals. These regions are then assigned higher mask values, while other regions are assigned lower mask values, thus forming a binary fused spectra mask. The fused spectra mask acts as an information filter, helping to identify and focus on time-frequency regions containing key signal features while ignoring areas with high noise, thereby improving signal purity and the accuracy of subsequent feature extraction. Furthermore, the fused spectra mask is used to enhance the fused spectra, ultimately recovering a purer target partial discharge sound signal. This can be achieved by multiplying the fused spectra with the corresponding fused spectra mask, thereby enhancing signal regions while suppressing noise regions. Performing an inverse short-time Fourier transform on the processed fused spectrum to recover the time-domain signal yields a continuous-time-domain partial discharge sound signal. This continuous-time-domain signal can be obtained as follows: ,in, This represents a continuous-time-domain partial discharge sound signal. The time-domain representation of any frame of signal. This represents the fused spectrum of any frame of the enhanced signal. This represents the fused spectral mask for any frame of signal. This represents the fused spectrum of any frame of the signal. Finally, the continuous time-domain partial discharge sound signals are superimposed according to a predetermined overlap rate and divided by the sum of squares of the window functions to compensate for the windowing effect, ultimately reconstructing the denoised continuous time-domain signal, which is the target partial discharge sound signal.

[0043] In one optional embodiment, determining the fused spectrum corresponding to each of the multiple frame signals based on the multiple target spectra corresponding to each of the multiple channels includes: determining the fused spectrum of any frame signal based on the multiple target spectra corresponding to each of the multiple channels in the following manner:

[0044] ;

[0045] in, This represents the fused spectrum of any frame of signal; This represents the target spectrum corresponding to any frame of signal in multiple spectrums of any channel; t represents the index of any frame of signal. Indicates the index of any channel. The total number of channels is represented by f; the frequency is represented by f; the fusion spectrum of each of the multiple frames of signals is obtained by using the method of obtaining the fusion spectrum of any frame signal.

[0046] Optionally, since partial discharge signals are present in multiple channels, while background noise distribution is relatively random, fusing the target spectra from different channels within the same frame can enhance shared signal features while weakening distinct noise features, thereby significantly improving the signal-to-noise ratio. Spectrum fusion ensures consistency of signals detected at the same time point across all channels, which is particularly important in multi-channel data processing, helping to eliminate signal inconsistencies caused by individual sensor failures or noise effects.

[0047] In one optional embodiment, determining the fusion spectrum mask corresponding to each of the multiple frame signals includes: determining the fusion spectrum mask of any frame signal by dividing the fusion spectrum of any frame signal into multiple sub-bands; determining the energy of multiple sub-bands of any frame signal, wherein the energy of each sub-band corresponds one-to-one with the energy of each sub-band; determining the energy change rate corresponding to each of the multiple sub-band energies; performing local peak detection on the energy of the multiple sub-bands to obtain the local peak results corresponding to each of the multiple sub-band energies; obtaining the fusion spectrum mask of any frame signal based on the energy change rate corresponding to each of the multiple sub-band energies and the local peak results corresponding to each of the multiple sub-band energies; and obtaining the fusion spectrum mask corresponding to each of the multiple frame signals by using the method of obtaining the fusion spectrum mask of any frame signal.

[0048] Optionally, firstly, the fused spectrum is divided into multiple sub-bands on the frequency axis. Each sub-band contains a continuous frequency range; this division allows for more refined analysis, focusing on the signal's characteristics in different frequency bands. For each sub-band of the fused spectrum of a frame of signal, its corresponding sub-band energy is calculated. The energy of any sub-band of any frame of signal can be determined as follows: ,in, Let represent the energy of any sub-band of any frame of signal, and b represent the index of any sub-band. Next, the rate of change of energy for each sub-band over time is calculated, which helps identify abrupt changes in signal strength and thus locate possible partial discharge events. The rate of change of energy for any sub-band can be obtained as follows: , This represents the rate of change of energy in any subband. , This represents the energy of any sub-band of the signal preceding any given frame. This indicates finding the minimum value. This indicates the absolute value. Further, local peak detection is performed on the energy of each sub-band to determine whether it reaches its maximum value at consecutive time points, thus filtering out moments of sudden increases in signal energy, as these moments are more likely to correspond to partial discharge events. The local peak value of any sub-band energy can be obtained as follows: , This represents the local peak energy result for any subband. This represents the energy of any sub-band of the signal in the next frame of any given frame. This indicates the search for the maximum value. Finally, based on the energy change rate of any subband energy and the results of local peak detection, a fused spectrogram mask for any frame of the signal is generated. The fused spectrogram mask for any frame of the signal can be obtained as follows: , This represents the fused spectral mask for any frame of signal. The masking mechanism automatically identifies and marks effective partial discharge information in the signal through energy change rate and local peak detection, effectively suppressing noise unrelated to discharge, especially noise that is widely distributed in the frequency domain and has unstable intensity.

[0049] Step S108: Determine the type of target partial discharge of the transformer based on the target partial discharge sound signal.

[0050] Optionally, based on the characteristics of the target partial discharge acoustic signal, the specific type of discharge can be accurately identified, including but not limited to corona discharge, surface discharge, internal discharge, and floating potential discharge. This is crucial for power system maintenance because partial discharge type identification helps power system maintenance personnel detect potential equipment faults at an early stage, preventing major accidents caused by insulation aging or damage. This is of great significance for ensuring power grid safety and preventing power grid failures.

[0051] In one optional embodiment, determining the target partial discharge type of the transformer based on the target partial discharge sound signal includes: performing multi-domain transformation on the target partial discharge sound signal to obtain a transformed partial discharge audio signal, wherein the multi-domain transformation includes frequency domain feature transformation and time-frequency domain feature transformation; obtaining multi-scale features of the target partial discharge audio signal based on the target partial discharge sound signal and the transformed partial discharge audio signal, wherein the multi-scale features represent the features of the target partial discharge audio signal obtained at multiple scales, and the multiple scales correspond one-to-one with multiple convolution kernel sizes; performing feature fusion on the multi-scale features to obtain fused features of the target partial discharge audio signal; and determining the target partial discharge type of the transformer based on the fused features.

[0052] Optionally, the denoised continuous time-domain signal, i.e., the target partial discharge sound signal, is first converted into frequency domain features and Mel spectrum. Frequency domain feature conversion can be obtained using short-time Fourier transform (SFT). This step converts the time-domain signal into a spectrum, showing the energy distribution of the signal at different frequency components, which is particularly beneficial for identifying high-frequency pulses. Mel spectrum conversion is obtained by filtering the SFT amplitude spectrum using a set of multiple Mel filter banks, focusing on the low-frequency details of the signal, which is more in line with human auditory perception characteristics, thus capturing subtle changes in the signal. This is crucial for identifying surface and internal discharges. Next, the features from the three different domains (time domain, frequency domain, and Mel spectrum) are fed into three convolutional branches of different scales for processing. For example, the small-scale branch (1×3 cores) specifically processes the time-domain signal, extracting detailed features of high-frequency signals such as corona discharge; the medium-scale branch (1×7 cores) analyzes the spectrum, extracting features of mid-frequency signals such as floating potential discharge; and the large-scale branch (1×11 cores) uses the Mel spectrum to capture global information of low-frequency signals such as surface and internal discharges. Furthermore, each branch can undergo two rounds of convolution and max pooling to obtain multi-scale features of the target partial discharge audio signal. The feature maps output from the multi-scale features are then concatenated along the channel dimension and batch normalized to obtain fused features. This fusion ensures a comprehensive understanding of the signal from multiple perspectives, covering the feature requirements of different discharge types. Next, the fused features can be split into time-domain and frequency-domain related feature parts, and the cosine similarity between them can be calculated to obtain a correlation matrix. Normalization is used to generate attention weights in the time and frequency domains, which reflect the importance of the time-domain and frequency-domain features to the overall signal. The time-domain and frequency-domain related features are then updated using time-domain and frequency-domain attention weights to enhance the most effective feature parts for discharge type identification. The updated features are then concatenated again to obtain the enhanced fused features. Finally, global average pooling is performed on the enhanced fused features to transform them into feature vectors. Subsequently, through fully connected layers and normalization functions, the classification probability distributions of various partial discharge types are obtained. The target discharge type is determined based on the highest probability category. A confidence threshold can be set; if the signal falls below this threshold, a mechanism to re-acquire the signal is triggered, thus improving the reliability of the classification. This process enables highly accurate and real-time identification of transformer partial discharge types, significantly enhancing the intelligence level of power system equipment condition monitoring. This is of great significance for preventing power equipment failures and ensuring the safe operation of the power grid.

[0053] Through the above steps S102 to S108, the target spectrum can be obtained by processing the acquired multi-channel partial discharge sound signal, and then the target partial discharge sound signal can be obtained, thereby accurately determining the target partial discharge type of the transformer. This achieves the technical effect of improving the accuracy of transformer partial discharge identification and solves the technical problem of insufficient accuracy of transformer partial discharge identification caused by incomplete consideration of factors when facing complex field environments.

[0054] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation method. Figure 2 This is a flowchart of an optional transformer partial discharge identification method according to an embodiment of the present invention, such as... Figure 2 As shown, the method includes:

[0055] S1: Data normalization. Bandpass filtering and normalization are performed on the partial discharge sound signals corresponding to each of the multiple channels in the acquired microphone array to obtain the normalized partial discharge sound signals corresponding to each of the multiple channels. The specific implementation process is the same as in the aforementioned embodiment, and will not be repeated here.

[0056] S2: Multi-channel short-time Fourier transform and target spectrum construction: The normalized partial discharge sound signal is converted into a time-frequency representation to obtain the target spectrum corresponding to each of the multiple channels. The specific implementation process is the same as the previous embodiment, and will not be repeated here.

[0057] S3: Multi-channel graph construction: Using multiple target spectra corresponding to each of the multiple channels, information fusion is performed in the time-frequency domain to determine the fused spectra corresponding to each of the multiple frame signals. The specific implementation process is the same as the aforementioned embodiment, and will not be repeated here.

[0058] S4: Subband pulse structure identification and initial mask generation. The fusion spectrum corresponding to each of the multiple frame signals is divided into multiple subbands. Local peak detection is performed on the energy of the multiple subbands to finally obtain the fusion spectrum mask corresponding to each of the multiple frame signals. The specific implementation process is the same as the previous embodiment, and will not be repeated here.

[0059] S5: Spectral mask filtering and reconstruction. The generated spectral mask is used to extract the continuous time-domain partial discharge sound signal from the fused spectrum, and the target partial discharge sound signal is finally reconstructed. The specific implementation process is the same as the previous embodiment, and will not be repeated here.

[0060] S6: Multi-domain conversion, converting the target partial discharge sound signal into frequency domain features and Mel spectrum. The specific implementation process is the same as the aforementioned embodiment, and will not be repeated here.

[0061] S7: Multi-scale convolution feature extraction. The obtained target partial discharge sound signal, frequency domain features and Mel spectrum features are respectively fed into three parallel convolution branches. The multi-scale features output by the three branches are concatenated in the channel dimension, and then batch normalized to output fused features. The specific implementation process is the same as the previous embodiment, and will not be repeated here.

[0062] S8: Cross-domain attention feature fusion, which splits the fused features into time-domain related parts and frequency-domain related parts. The specific implementation process is the same as the previous embodiment, and will not be repeated here.

[0063] S9: State recognition. Through global average pooling, fully connected layers, and normalization calculation, the probability distribution of various partial discharge types is obtained. The category with the highest probability is taken as the target partial discharge type. The specific implementation process is the same as the aforementioned embodiments, and will not be repeated here.

[0064] This embodiment also provides a partial discharge identification device for a transformer, which is used to implement the above embodiments and preferred embodiments; details already described will not be repeated. As used below, the terms "module" and "device" can refer to a combination of software and / or hardware that performs a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, hardware implementation, or a combination of software and hardware, is also possible and contemplated.

[0065] According to an embodiment of the present invention, an apparatus embodiment for implementing the above-described transformer partial discharge identification method is also provided. Figure 3 This is a schematic diagram of the structure of a partial discharge identification device for a transformer according to an embodiment of the present invention, as shown below. Figure 3 As shown, the partial discharge identification device for the above-mentioned transformer includes: a partial discharge sound signal acquisition module 300, a target spectrum determination module 302, a target partial discharge sound signal determination module 304, and a target partial discharge type determination module 306, wherein:

[0066] The partial discharge sound signal acquisition module 300 is used to acquire the partial discharge sound signals corresponding to each of the multiple channels in the microphone array, wherein the partial discharge sound signals are acquired from the transformer;

[0067] The target spectrum determination module 302 is connected to the partial discharge sound signal acquisition module 300. It is used to obtain the target spectrum corresponding to each of the multiple channels based on the partial discharge sound signals corresponding to each channel. The target spectrum represents the energy distribution of the corresponding partial discharge sound signal at multiple frequencies and multiple times.

[0068] The target partial discharge sound signal determination module 304 is connected to the target spectrum determination module 302 and is used to obtain the target partial discharge sound signal based on the target spectrum corresponding to each of the multiple channels.

[0069] The target partial discharge type determination module 306 is connected to the target partial discharge sound signal determination module 304 and is used to determine the target partial discharge type of the transformer based on the target partial discharge sound signal.

[0070] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0071] It should be noted that the aforementioned partial discharge sound signal acquisition module 300, target spectrum determination module 302, target partial discharge sound signal determination module 304, and target partial discharge type determination module 306 correspond to steps S102 to S108 in the embodiments. The instances and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the content disclosed in the above embodiments. It should be noted that the above modules, as part of the device, can run in a computer terminal.

[0072] It should be noted that the optional or preferred implementation methods of this embodiment can be found in the relevant descriptions in the embodiments, and will not be repeated here.

[0073] The aforementioned transformer partial discharge identification device may further include a processor and a memory. The aforementioned partial discharge sound signal acquisition module 300, target spectrum determination module 302, target partial discharge sound signal determination module 304, target partial discharge type determination module 306, etc., are all stored in the memory as program modules, and the processor executes the aforementioned program modules stored in the memory to realize the corresponding functions.

[0074] The processor contains a core that retrieves the corresponding program modules from memory. One or more cores may be configured. Memory may include non-persistent memory in computer-readable media, such as random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. Memory includes at least one memory chip.

[0075] According to an embodiment of this application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein, when the program runs, it controls the device containing the non-volatile storage medium to execute any of the aforementioned transformer partial discharge identification methods.

[0076] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a group of computer terminals in a computer network, or in any mobile terminal in a group of mobile terminals, and the non-volatile storage medium includes stored programs.

[0077] Optionally, during program execution, the device containing the non-volatile storage medium may be controlled to perform the following functions: acquire partial discharge sound signals corresponding to multiple channels in the microphone array, wherein the partial discharge sound signals are acquired from the transformer; based on the partial discharge sound signals corresponding to multiple channels, obtain target spectra corresponding to multiple channels, wherein the target spectra represent the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times; based on the target spectra corresponding to multiple channels, obtain target partial discharge sound signals; and based on the target partial discharge sound signals, determine the target partial discharge type of the transformer.

[0078] According to an embodiment of this application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein the program executes any of the aforementioned transformer partial discharge identification methods.

[0079] According to an embodiment of this application, an embodiment of a computer program product is also provided. Optionally, in this embodiment, the computer program product includes a computer program that, when executed by a processor, implements the steps of the partial discharge identification method for transformers described above.

[0080] Optionally, when the above-mentioned computer program product is executed on a data processing device, it is suitable to execute an initialization program having the following method steps: acquiring partial discharge sound signals corresponding to multiple channels in a microphone array, wherein the partial discharge sound signals are acquired from a transformer; obtaining target spectra corresponding to multiple channels based on the partial discharge sound signals corresponding to each channel, wherein the target spectra represent the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times; obtaining target partial discharge sound signals based on the target spectra corresponding to each channel; and determining the target partial discharge type of the transformer based on the target partial discharge sound signals.

[0081] This invention provides an electronic device, which includes a processor, a storage unit, and a storage unit for acquiring partial discharge sound signals corresponding to multiple channels in a microphone array, wherein the partial discharge sound signals are acquired from a transformer. Based on the partial discharge sound signals corresponding to the multiple channels, target spectrograms corresponding to the multiple channels are obtained, wherein the target spectrograms represent the energy distribution of the corresponding partial discharge sound signals at multiple frequencies and multiple times. Based on the target spectrograms corresponding to the multiple channels, target partial discharge sound signals are obtained. Based on the target partial discharge sound signals, the target partial discharge type of the transformer is determined.

[0082] The order of the above embodiments of the present invention is merely for description and does not represent the superiority or inferiority of the embodiments.

[0083] In the above embodiments of the present invention, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0084] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. The device embodiments described above are merely illustrative; for example, the division of modules described above can be a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces, or indirect coupling or communication connection between modules, and may be electrical or other forms.

[0085] The modules described above as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0086] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated modules described above can be implemented in hardware or as software functional modules.

[0087] If the aforementioned integrated modules are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a non-volatile storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned non-volatile storage medium includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0088] The above are merely preferred embodiments of the present invention. It should be noted that those skilled in the art can make various improvements and modifications without departing from the principle of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A method for identifying partial discharge in a transformer, characterized in that, include: Acquire the partial discharge sound signals corresponding to each of the multiple channels in the microphone array, wherein the partial discharge sound signals are acquired by a transformer; Based on the partial discharge sound signals corresponding to each of the multiple channels, a target spectrum corresponding to each of the multiple channels is obtained, wherein the target spectrum represents the energy distribution of the corresponding partial discharge sound signal at multiple frequencies and multiple times; Based on the target spectrum corresponding to each of the multiple channels, the target partial discharge sound signal is obtained; Based on the target partial discharge sound signal, the target partial discharge type of the transformer is determined.

2. The method according to claim 1, characterized in that, The process of obtaining the target spectrum corresponding to each of the multiple channels based on the partial discharge sound signals corresponding to each channel includes: Bandpass filtering is performed on the partial discharge sound signal of any channel to obtain the filtered partial discharge sound signal corresponding to any channel. The filtered partial discharge sound signal is normalized to obtain the normalized partial discharge sound signal corresponding to any channel. Based on the normalized partial discharge sound signal, the target spectrum of any channel is obtained; Based on the partial discharge sound signals corresponding to each of the multiple channels, the target spectrum corresponding to each of the multiple channels is obtained by using the method of obtaining the target spectrum of any one of the channels.

3. The method according to claim 2, characterized in that, When there are multiple target spectra for any channel, obtaining the target spectra for any channel based on the normalized partial discharge sound signal includes: The normalized partial discharge sound signal is subjected to frame segmentation processing to obtain multiple frame signals for any channel, wherein the frame signal represents a signal segment of a predetermined sampling duration obtained from the normalized partial discharge sound signal. Determine the target spectrogram corresponding to each of the multiple frame signals; Based on the target spectra corresponding to the multiple frame signals of any one channel, multiple target spectra of any one channel are obtained.

4. The method according to claim 3, characterized in that, The step of performing frame segmentation processing on the normalized partial discharge sound signal to obtain multiple frame signals for any channel includes: The normalized partial discharge sound signal is processed by framing to obtain any frame signal of any channel: ； in, This represents any frame signal of any channel; i represents the index of any channel. The index of any frame signal is represented by L, where L represents the number of frame signals; This represents the index of a sampling point in any frame of the signal. This indicates the predetermined sampling duration; H represents the normalized partial discharge sound signal; H represents the frame shift between the starting sampling points of two adjacent frame signals. Multiple frame signals of any channel are obtained by obtaining any frame signal of any channel.

5. The method according to claim 3, characterized in that, Determining the target spectral map corresponding to each of the plurality of frame signals includes: The target spectrum of any frame of signal is determined as follows: ； in, This represents the target spectrum of any frame of the signal; f represents the frequency; ; Indicates the index of any of the aforementioned frame signals. This represents the short-time Fourier transform result of any frame of the signal; Represents the Hanning window function; By obtaining the target spectrum of any frame signal, the target spectrum of each of the multiple frame signals of any channel is obtained.

6. The method according to claim 3, characterized in that, When there are multiple target spectra corresponding to each channel, and multiple frame signals correspond one-to-one with multiple target spectra, the step of obtaining the target partial discharge sound signal based on the target spectra corresponding to each of the multiple channels includes: Based on the multiple target spectra corresponding to each of the multiple channels, a fused spectra corresponding to each of the multiple frame signals is determined, wherein the fused spectra is obtained by fusing the target spectra corresponding to the same frame signal from the multiple target spectra corresponding to each of the multiple channels; Determine the fusion spectrogram mask corresponding to each of the multiple frame signals; Based on the fusion spectrum corresponding to each of the multiple frame signals, and the fusion spectrum mask corresponding to each of the multiple frame signals, the target partial discharge sound signal is obtained.

7. The method according to claim 6, characterized in that, The step of determining the fused spectrum corresponding to each of the multiple frame signals based on the multiple target spectrums corresponding to each of the multiple channels includes: Based on the multiple target spectra corresponding to each of the multiple channels, the fused spectra of any frame signal are determined in the following manner: ； in, This represents the fused spectrum of any one of the frame signals; This represents the target spectrum corresponding to any frame signal in multiple spectrums of any channel; t represents the index of any frame signal. Indicates the index of any of the channels. The total number of channels is indicated by f; f represents the frequency. By obtaining the fusion spectrum of any frame signal, the fusion spectrum corresponding to each of the multiple frame signals is obtained.

8. The method according to claim 6, characterized in that, Determining the fused spectrogram mask corresponding to each of the plurality of frame signals includes: The fused spectral mask for any frame of signal is determined as follows: The fused spectrum of any frame signal is divided into multiple sub-bands; Determine the energy of multiple sub-bands for any given frame signal, wherein the energy of each sub-band corresponds one-to-one with the energy of the multiple sub-bands; Determine the rate of change of energy corresponding to each of the multiple sub-band energies; Local peak detection is performed on the multiple sub-band energies to obtain the local peak results corresponding to each of the multiple sub-band energies; Based on the energy change rate corresponding to each of the multiple sub-band energies and the local peak results corresponding to each of the multiple sub-band energies, the fused spectral mask of any frame signal is obtained; By obtaining the fusion spectrum mask of any one of the frame signals, the fusion spectrum mask corresponding to each of the multiple frame signals is obtained.

9. The method according to any one of claims 1 to 8, characterized in that, The determination of the target partial discharge type of the transformer based on the target partial discharge sound signal includes: The target partial discharge sound signal is subjected to multi-domain transformation to obtain the transformed partial discharge audio signal, wherein the multi-domain transformation includes frequency domain feature transformation and time-frequency domain feature transformation; Based on the target partial discharge sound signal and the converted partial discharge audio signal, multi-scale features of the target partial discharge audio signal are obtained, wherein the multi-scale features represent the features of the target partial discharge audio signal obtained at multiple scales, and the multiple scales correspond one-to-one with multiple convolution kernel sizes; The multi-scale features are fused to obtain the fused features of the target partial discharge audio signal; Based on the fusion characteristics, the target partial discharge type of the transformer is determined.

10. An electronic device, characterized in that, It includes one or more processors and a memory, the memory being used to store one or more programs for execution, wherein when the one or more programs are executed by the one or more processors, the one or more processors cause the one or more processors to implement the partial discharge identification method for a transformer as described in any one of claims 1 to 9.

Images