Method and device for determining partial discharge defect type of gas insulated switchgear

Abstract

The invention discloses a method and a device for determining a partial discharge defect type of gas insulated switchgear. The method comprises the following steps: acquiring a partial discharge pulse sequence of a target gas insulated switchgear in a target period; based on the partial discharge pulse sequence, determining a phase resolution partial discharge matrix of the target gas insulated switchgear in the target period; based on the phase resolution partial discharge matrix, determining a target feature vector of the target gas insulated switchgear in the target period; and determining a target partial discharge defect type of the target gas insulated switchgear in the target period based on the target feature vector. According to the method and the device, the technical problem that the partial discharge defect type determination result of the gas insulated switchgear is inaccurate in the prior art is solved.

Description

Technical Field

[0001] This application relates to the field of power systems, and more specifically, to a method and apparatus for determining the type of partial discharge defects in gas-insulated switchgear. Background Technology

[0002] As a key component of power grids, the insulation status monitoring of GIS (Gas Insulated Switchgear) is crucial. In partial discharge (PD) detection, the Phase-Resolved Partial Discharge Pattern (PRPD) is the most critical diagnostic basis. Currently, most AI-based PD fault diagnosis systems use PRPD maps as input. However, a standard PRPD map relies on accurate power frequency voltage phase as the horizontal axis reference. There are two methods to obtain the power frequency voltage phase: one is to acquire the phase of the PT (Potential Transformer) signal on-site, but this carries a short-circuit risk and is extremely dangerous; the other is non-contact synchronization (such as UHF external antennas or electric field sensors), but this is susceptible to stray capacitance, causing random phase shifts in the acquired PRPD map.

[0003] To address the issue of phase shift in PRPD maps, related techniques artificially augment the dataset during training by performing numerous random shifts on the PRPD map. This method, however, leads to convergence difficulties and the model doesn't truly understand "phase independence," merely memorizing all positions, resulting in high parameter redundancy. While convolutional neural networks possess some translation invariance, they remain extremely sensitive to large-scale global periodic phase shifts (e.g., 180-degree shifts) due to the limited stride of the pooling layers. This causes a sharp drop in the accuracy of diagnostic systems recognizing PRPD map data with phase shifts. Therefore, related techniques suffer from inaccurate determination of partial discharge defect types in gas-insulated switchgear due to phase shifts in PRPD maps.

[0004] There is currently no effective solution to the above problems. Summary of the Invention

[0005] This application provides a method and apparatus for determining the type of partial discharge defect in gas-insulated switchgear, so as to at least solve the technical problem of inaccurate determination of the type of partial discharge defect in gas-insulated switchgear in related technologies.

[0006] According to one aspect of the embodiments of this application, a method for determining the partial discharge defect type of a gas-insulated switchgear is provided, comprising: acquiring a partial discharge pulse sequence of a target gas-insulated switchgear in a target period; determining a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulses, the columns represent the phases of the partial discharge pulses, and the phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulses of the target gas-insulated switchgear in the target period; determining a target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix; and determining the target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector.

[0007] According to another aspect of the embodiments of this application, a device for determining the partial discharge defect type of a gas-insulated switchgear is provided, comprising: a data acquisition module for acquiring a partial discharge pulse sequence of a target gas-insulated switchgear in a target period; a first determination module for determining a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulses, the columns represent the phases of the partial discharge pulses, and the phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulses of the target gas-insulated switchgear in the target period; a second determination module for determining a target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix; and a third determination module for determining the target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector.

[0008] According to another aspect of the embodiments of this application, a non-volatile storage medium is provided, which stores multiple instructions adapted for a method for determining the partial discharge defect type of a gas-insulated switchgear, any one of which is loaded and executed by a processor.

[0009] According to another aspect of the embodiments of this application, an electronic device is 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 any one of the following methods for determining the partial discharge defect type of a gas-insulated switchgear.

[0010] According to another aspect of the embodiments of this application, a computer program product is provided, which, when executed on a data processing device, is adapted to perform the steps of a method for determining the type of partial discharge defect in a gas-insulated switchgear.

[0011] In this embodiment, the partial discharge pulse sequence of the target gas-insulated switchgear in a target period is obtained; based on the partial discharge pulse sequence, a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period is determined, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulses, and the columns represent the phases of the partial discharge pulses. The phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulses of the target gas-insulated switchgear in the target period; based on the phase-resolved partial discharge matrix, a target feature vector of the target gas-insulated switchgear in the target period is determined; and based on the target feature vector, the target partial discharge defect type of the target gas-insulated switchgear in the target period is determined. This achieves the goal of determining the phase-resolved partial discharge matrix based on the obtained partial discharge pulse sequence of the target gas-insulated switchgear, thereby obtaining the target partial discharge defect type of the target gas-insulated switchgear. This improves the accuracy of the determination of the target partial discharge defect type of the target gas-insulated switchgear, thus solving the technical problem of inaccurate determination of the partial discharge defect type of gas-insulated switchgear in related technologies. Attached Figure Description

[0012] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0013] Figure 1 This is a flowchart of a method for determining the type of partial discharge defect in a gas-insulated switchgear according to an embodiment of this application;

[0014] Figure 2 This is a flowchart of an optional method for determining the type of partial discharge defect in a gas-insulated switchgear according to an embodiment of this application;

[0015] Figure 3 This is a flowchart of an optional method for determining a target feature vector according to an embodiment of this application;

[0016] Figure 4 This is a schematic diagram of an optional device for determining the partial discharge defect type of a gas-insulated switchgear according to an embodiment of this application. Detailed Implementation

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

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application 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 this application 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 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] It should be noted that the information and data collected in this application (including but not limited to partial discharge pulse sequences) are information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of this data all comply with relevant laws, regulations, and standards, and necessary confidentiality measures have been taken. This process does not violate public order and good morals, and corresponding access points are provided for users to choose whether to authorize or refuse. For example, interfaces are set up between this system and relevant users or organizations, providing users with corresponding access points to choose whether to agree to or refuse the automated decision-making results; if the user chooses to refuse, the process proceeds to the expert decision-making stage.

[0020] According to an embodiment of this application, a method embodiment for determining the partial discharge defect type of a gas-insulated switchgear 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.

[0021] Figure 1 This is a flowchart of a method for determining the type of partial discharge defect in a gas-insulated switchgear according to an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps:

[0022] Step S102: Obtain the partial discharge pulse sequence of the target gas-insulated switchgear in the target cycle;

[0023] It is understandable that by acquiring the partial discharge pulse sequence within the target period with high precision under conditions without external synchronization, the reliability of frequency-space dual-domain feature extraction can be improved, thereby increasing the accuracy of the determination of the target partial discharge defect type of the target gas-insulated switchgear.

[0024] Alternatively, the partial discharge pulse sequence inside the target gas-insulated switchgear can be acquired using an ultra-high frequency (UHF) sensor or other partial discharge sensor.

[0025] Step S104: Based on the partial discharge pulse sequence, determine the phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period. The rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulse, and the columns represent the phase of the partial discharge pulse. The phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulse of the target gas-insulated switchgear in the target period.

[0026] It is understandable that, based on the collected partial discharge pulse sequence, a phase-resolved partial discharge matrix (PRPD) of the target gas-insulated switchgear in the target period is constructed, i.e., a PRPD map represented in matrix form. The phase-resolved partial discharge matrix describes the amplitude and phase distribution characteristics of the partial discharge pulses in the target gas-insulated switchgear during the target period; rows represent the amplitude of the partial discharge pulses, and columns represent the phase. By converting the partial discharge pulse sequence into a phase-resolved partial discharge matrix, the microscopic morphological distribution of partial discharges in the amplitude-phase space is fully preserved, providing high-fidelity input for subsequent deep learning models and improving the identification accuracy of typical partial discharge defect types such as insulator air gaps, floating potentials, and free particles.

[0027] Optionally, based on the acquired partial discharge pulse sequence inside the target gas-insulated switchgear, a phase-resolved partial discharge matrix can be constructed as follows: First, set the power frequency period to T (usually 20 ms), and divide it into M phase windows (e.g., M = 360 degrees, i.e., 1 degree per window). Second, set the amplitude range of the partial discharge pulses, and divide it into N amplitude levels (e.g., N = 128). Finally, traverse all acquired partial discharge pulse sequences, and for each pulse, accumulate it into the corresponding cell (n, m) in the PRPD matrix according to its corresponding power frequency phase angle and amplitude, where n ∈ [1, N] represents the amplitude level and m ∈ [1, M] represents the phase. The element values ​​of the obtained phase-resolved partial discharge matrix characterize the cumulative frequency or energy intensity of the partial discharge pulses under a specific amplitude-phase combination, and are used to describe the distribution characteristics of partial discharge in the amplitude-phase two-dimensional space.

[0028] Step S106: Based on the phase-resolved partial discharge matrix, determine the target feature vector of the target gas-insulated switchgear in the target period;

[0029] It is understandable that the target feature vector obtained from the constructed phase-resolved partial discharge matrix can eliminate phase drift interference and improve the accuracy of the determination of the target partial discharge defect type of the target gas-insulated switchgear.

[0030] In one optional embodiment, determining the target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix includes: normalizing the phase-resolved partial discharge matrix to obtain a standardized matrix; reconstructing the standardized matrix to obtain a reconstructed matrix, wherein each element in the reconstructed matrix is ​​a real number used to characterize the intensity of the partial discharge defect at the corresponding amplitude level; extracting features from the reconstructed matrix to obtain a first feature vector of the target gas-insulated switchgear in the target period; extracting features from the standardized matrix to obtain a second feature vector of the target gas-insulated switchgear in the target period; and determining the target feature vector based on the first and second feature vectors.

[0031] The target eigenvector of the gas-insulated switchgear in the target period can be determined as follows: First, the element values ​​in the phase-resolved partial discharge matrix are normalized to obtain a standardized matrix. Second, the standardized matrix is ​​reconstructed to obtain a reconstructed matrix. Then, features are extracted from the reconstructed matrix to obtain the first eigenvector of the gas-insulated switchgear in the target period, and features are extracted from the standardized matrix to obtain the second eigenvector of the gas-insulated switchgear in the target period. Finally, the first and second eigenvectors are fused to obtain the target eigenvector. This method can eliminate partial discharge diagnostic failures caused by phase drift, achieve decoupling of the "morphology-period" dual-dimensional features, improve the rationality and accuracy of the target eigenvector, and thus improve the accuracy of the determination of the target partial discharge defect type of the gas-insulated switchgear.

[0032] Optionally, to accelerate model convergence, the phase-resolved partial discharge matrix X can be subjected to max-min normalization, mapping the element values ​​in the matrix to the interval [0,1] to obtain the normalized matrix X. norm (i.e., the standardized PRPD map represented in matrix form).

[0033] Optionally, the AF-Map (i.e., the reconstruction matrix) can be input into the first convolutional neural network branch. This branch uses the ResNet-18 architecture and extracts high-dimensional frequency domain feature vectors (i.e., the first feature vector) through convolution and pooling operations. Although the phase has shifted, the local geometric textures of the partial discharge PRPD pattern corresponding to the normalized matrix (such as the "rabbit ear" shape of air gap discharge and the "triangular" shape of suspension discharge) still contain important partial discharge information. Therefore, feature extraction is performed on the normalized matrix to obtain the second eigenvector. Thus, the normalized matrix X can be... norm The data is directly input into the second convolutional neural network branch, which has a similar or identical structure to the first branch and is used to extract the spatial texture feature vector (i.e., the second feature vector). The above process ensures that the model does not lose the intuitive morphological features in the normalized matrix due to excessive focus on frequency domain information.

[0034] In one optional embodiment, the normalized matrix is ​​reconstructed to obtain a reconstructed matrix, including: performing a discrete Fourier transform on the normalized matrix to obtain a frequency domain matrix, wherein each element in the frequency domain matrix is ​​a complex number used to characterize the amplitude and phase information of the harmonic components contained in the partial discharge pulse within the target period at the corresponding amplitude level; and extracting the magnitude of each element in the frequency domain matrix to obtain the reconstructed matrix.

[0035] It can be understood that performing a one-dimensional discrete Fourier transform on the normalized matrix along the phase axis yields a frequency domain matrix, where each element is a complex number representing the harmonic amplitude and phase information of the partial discharge pulse within the power frequency cycle at a specific amplitude level. Extracting the magnitude of each complex element in the frequency domain matrix yields a reconstructed matrix. The elements of this reconstructed matrix only represent the energy distribution, do not contain phase information, and are invariant to any translation along the phase axis. Through this method, phase drift interference can be eliminated, achieving "zero-synchronization" partial discharge diagnosis, suppressing noise and random interference, enhancing feature robustness, and the resulting reconstructed matrix can reveal the physical nature of the discharge, enhancing feature interpretability.

[0036] Alternatively, the normalized matrix X can be... norm For each row, perform a one-dimensional discrete Fourier transform along the phase axis, as shown in the following equation:

[0037]

[0038] in, The normalized discharge intensity of the nth amplitude level and the mth phase dimension is represented by the normalized matrix X. norm The value of the element in the nth row and mth column, where j is the imaginary unit and k is the frequency component. It represents the complex spectral value of the nth amplitude level at frequency component k, which is the element value in the nth row and kth column of the frequency domain matrix.

[0039] Alternatively, the magnitude of the discrete Fourier transform result can be calculated, discarding the phase spectrum information, as shown in the following equation:

[0040]

[0041] in, This represents the amplitude spectrum value of the nth amplitude level at frequency component k, which is the element value in the nth row and kth column of the reconstruction matrix.

[0042] According to the time-shift property of Fourier transform, if the partial discharge pulse signal undergoes a phase shift in the time domain... Its frequency domain results This will generate a size of Phase rotation (i.e., multiplying the frequency domain result by) ), and module length It remains constant. Therefore, It is a physical quantity that does not change with phase drift.

[0043] Alternatively, the normalized matrix can be reconstructed in the following manner. For the normalized matrix X... normFor each row (i.e., each amplitude level n), perform a one-dimensional discrete Fourier transform along the phase axis (M columns in total) and calculate its frequency domain complex result; retain the amplitude spectrum (i.e., magnitude). Discarding the phase spectrum, we obtain the amplitude vector corresponding to each amplitude level. Stack all N amplitude vectors in their original order to form a two-dimensional matrix, defined as the amplitude-frequency matrix (AF-Map), where each element... is a real number, representing the periodic energy distribution intensity of the partial discharge pulse at amplitude level n and frequency component k.

[0044] In one optional embodiment, determining a target feature vector based on a first feature vector and a second feature vector includes: concatenating the first feature vector and the second feature vector to obtain a concatenated feature vector; determining an initial weight of the first feature vector based on the concatenated feature vector; correcting the initial weight to obtain a target weight of the first feature vector; and fusing the first feature vector and the second feature vector based on the target weight to obtain the target feature vector.

[0045] The target feature vector is obtained by using the following method based on the first and second feature vectors. First, the first and second feature vectors are concatenated to obtain a concatenated feature vector. Second, the initial weight of the first feature vector is obtained based on the concatenated feature vector. Then, the initial weight is corrected to obtain the target weight of the first feature vector. Finally, the first and second feature vectors are weighted and fused based on the target weight to obtain the final target feature vector. By correcting the initial weight to obtain the target weight, the weighted fusion of the first and second feature vectors is achieved, enabling the partial discharge diagnosis process to stably output high-precision diagnostic results even in complex field environments with drastic phase drift.

[0046] Optionally, to enable the model to automatically determine whether the current phase drift is severe and dynamically adjust the weights of the two branches, an adaptive gating fusion mechanism can be used to fuse the first and second feature vectors. First, the high-dimensional frequency domain feature vectors are... Spatial texture feature vector First, concatenate the features along the channel dimension to obtain a concatenated feature vector. Then, input the concatenated feature vector into a fully connected layer (FC), pass it through a sigmoid activation function, and output a scalar weight. (i.e., initial weights) This represents the model's "degree of trust" in the high-dimensional frequency domain feature vector. Finally, the final diagnostic feature vector (i.e., the target feature vector) is calculated. .

[0047]

[0048] In one optional embodiment, the initial weights are corrected to obtain the target weights of the first eigenvector, including: determining the frequency domain energy concentration based on the first eigenvector, wherein the frequency domain energy concentration is used to characterize the degree of aggregation of the power frequency fundamental wave and low-order harmonic components in the reconstructed matrix; determining the spatial domain texture disorder based on the second eigenvector, wherein the spatial domain texture disorder is used to quantify the average gray-level entropy of the normalized matrix, and the average gray-level entropy is used to describe the degree of local structural blurring of the normalized matrix caused by phase drift or noise interference; and correcting the initial weights based on the frequency domain energy concentration and the spatial domain texture disorder to obtain the target weights.

[0049] It can be understood that, based on the first eigenvector, the frequency domain energy concentration, used to characterize the degree of aggregation of the fundamental and low-order harmonic components in the reconstructed matrix, is determined. Simultaneously, based on the second eigenvector, the spatial domain texture disorder, used to quantify the average gray-level entropy of the normalized matrix, is determined. The initial weights are then corrected based on the frequency domain energy concentration and spatial domain texture disorder to obtain the target weights. This weight correction based on the frequency domain energy concentration and spatial domain texture disorder enables proactive sensing and adaptive compensation of phase drift interference, improving the accuracy of the target weight determination results.

[0050] Optionally, the rows of the normalized matrix represent the amplitude, and the columns represent the phase. The rows of the reconstructed matrix represent the amplitude, and the columns represent the frequency components.

[0051] Alternatively, the frequency domain energy concentration can be determined in the following manner. .

[0052]

[0053]

[0054] in, To reconstruct the energy of the fundamental power frequency wave in the nth amplitude level (i.e., the nth row) of the matrix, and The energy of the lower harmonics in the nth amplitude level. The sum of the energy of the fundamental frequency wave and the lower harmonics in the nth amplitude level is given by N, where N is the total number of amplitude levels.

[0055] Alternatively, the spatial texture disorder ATC can be determined in the following manner.

[0056]

[0057] Where M is the number of columns in the normalized matrix, Let be the gray-level entropy of the m-th phase dimension in the normalization matrix. The smaller the average gray-level entropy, the clearer the texture of the PRPD image corresponding to the normalization matrix; the larger the average gray-level entropy, the more severe the blurring caused by phase drift or noise in the PRPD image corresponding to the normalization matrix.

[0058] Alternatively, the target weights can be determined in the following manner.

[0059]

[0060] in, As the initial weights, and This is the normalization coefficient.

[0061] Step S108: Based on the target feature vector, determine the type of target partial discharge defect in the target gas-insulated switchgear during the target period.

[0062] It is understandable that by inputting the target feature vector into the trained classifier, the type of partial discharge defect of the target gas-insulated switchgear in the target period can be obtained.

[0063] In one optional embodiment, determining the target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector includes: determining the probabilities corresponding to various candidate partial discharge defect types of the target gas-insulated switchgear in the target period based on the target feature vector; and determining the candidate partial discharge defect type corresponding to the highest probability among the probabilities corresponding to the various candidate partial discharge defect types as the target partial discharge defect type.

[0064] It can be understood that by inputting the target feature vector into a trained classifier, the probabilities of various candidate partial discharge defect types for the target gas-insulated switchgear during the target period are obtained. These candidate partial discharge defect types may include, but are not limited to, metal spikes, insulator air gaps, floating potentials, and free particles. The candidate partial discharge defect type with the highest probability among these probabilities is determined as the target partial discharge defect type for the target gas-insulated switchgear during the target period. By accurately determining the target features and combining them with the trained classifier, the accuracy of the probability determination of candidate partial discharge defect types is improved, thereby improving the accuracy of the target partial discharge defect type determination.

[0065] Optionally, the fused target feature vector can be... Input the trained classifier (typically composed of fully connected layers and Softmax layers). Output the probability distribution of various candidate partial discharge defect types (such as metal spikes, insulator air gaps, floating potentials, free particles, normal operation, etc.). Select the candidate partial discharge defect type with the highest probability as the final diagnostic result (i.e., the target partial discharge defect type).

[0066] Through the above steps S102 to S108, the phase-resolved partial discharge matrix can be determined based on the obtained partial discharge pulse sequence of the target gas-insulated switchgear, thereby obtaining the target partial discharge defect type of the target gas-insulated switchgear. This achieves the technical effect of improving the accuracy of the determination result of the target partial discharge defect type of the target gas-insulated switchgear, and solves the technical problem of inaccurate determination result of the partial discharge defect type of gas-insulated switchgear in related technologies.

[0067] Based on the above embodiments and optional embodiments, this application proposes an implementation method for determining the partial discharge defect type of gas-insulated switchgear. This implementation method can be understood as a robust diagnostic method for partial discharge in GIS based on frequency-space dual-domain feature decoupling. By starting from the physical essence of signal processing and utilizing the translation invariance of the amplitude spectrum of Discrete Fourier Transform (DFT), a dual-stream fusion network of "frequency domain invariant + spatial domain texture" is constructed. This ensures that extremely high fault diagnosis accuracy can be maintained under arbitrary phase drift, solving the problem that the GIS partial discharge diagnosis in the prior art relies heavily on precise synchronization signals and that the diagnostic accuracy decreases due to phase drift. This achieves high-precision partial discharge fault type diagnosis of gas-insulated switchgear under phase synchronization loss conditions.

[0068] Figure 2 This is a flowchart of an optional method for determining the type of partial discharge defect in a gas-insulated switchgear according to an embodiment of this application, such as... Figure 2 As shown, the steps of the robust diagnostic method for partial discharge in GIS based on frequency-space dual-domain feature decoupling include:

[0069] Step S1: Construction and standardization of the PRPD matrix (i.e., phase-resolved partial discharge matrix).

[0070] Partial discharge pulse sequences inside the target gas-insulated switchgear are acquired using an Ultra High Frequency (UHF) sensor or other partial discharge sensor. These sequences are then mapped to a phase-resolved partial discharge matrix (i.e., a PRPD spectrum represented in matrix form). The steps include:

[0071] Set the power frequency period to T (usually 20ms), and divide it into M phase windows (e.g., M=360 degrees, i.e., 1 degree per window).

[0072] Define the amplitude range of the partial discharge pulses and divide it into N amplitude levels (e.g., N=128). Statistically count the cumulative amplitude or frequency of the partial discharge pulses falling into each matrix unit (n,m) in the partial discharge pulse sequence collected per unit time, and construct the original PRPD spectrum matrix X (i.e., the phase-resolved partial discharge matrix, which is an n×m matrix).

[0073] To accelerate model convergence, matrix X is subjected to min-max normalization, mapping the element values ​​to the interval [0,1] to obtain the normalized matrix X. norm (i.e., the standardized PRPD map represented in matrix form).

[0074] Step S2: Construct the "frequency domain translation invariant" feature extraction branch.

[0075] This step aims to extract frequency domain features that are “naturally immune” to phase drift.

[0076] Step S21, one-dimensional discrete Fourier transform of the phase axis.

[0077] For the standardized matrix X norm For each row, perform a one-dimensional discrete Fourier transform along the phase axis, as shown in the following equation:

[0078]

[0079] in, The normalized discharge intensity of the nth amplitude level and the mth phase dimension is represented by the normalized matrix X. norm The value of the element in the nth row and mth column, where j is the imaginary unit and k is the frequency component. It represents the complex spectral value of the nth amplitude level at frequency component k, which is the element value in the nth row and kth column of the frequency domain matrix.

[0080] Step S22, amplitude spectrum extraction.

[0081] Calculate the magnitude of the discrete Fourier transform result, discarding the phase spectrum information, as follows:

[0082]

[0083] in, This represents the amplitude spectrum value of the nth amplitude level at frequency component k, which is the element value in the nth row and kth column of the reconstruction matrix.

[0084] According to the time-shift property of Fourier transform, if the partial discharge pulse signal undergoes a phase shift in the time domain... Its frequency domain results This will generate a size of Phase rotation (i.e., multiplying the frequency domain result by) ), and module length It remains constant. Therefore, It is a physical quantity that does not change with phase drift.

[0085] Step S23, amplitude-frequency feature map reconstruction.

[0086] Calculated from all N rows The components are re-stacked in their original order to generate a new two-dimensional matrix, defined as the amplitude-frequency matrix AF-Map (i.e., the reconstruction matrix). This reconstruction matrix reflects the periodic distribution density of the partial discharge signal at different amplitude levels and is completely unaffected by the initial phase.

[0087] Step S24, frequency domain deep feature encoding.

[0088] The AF-Map is input into the first convolutional neural network branch. This branch uses the ResNet-18 architecture and extracts high-dimensional frequency domain feature vectors (i.e., the first feature vector) through convolution and pooling operations. .

[0089] Step S3: Construct the "spatial morphology texture" feature extraction branch.

[0090] Despite the phase shift, the local geometric textures of the partial discharge PRPD map corresponding to the normalized matrix (such as the "rabbit ear" shape of air gap discharge and the "triangular" shape of suspension discharge) still contain important partial discharge information. Therefore, feature extraction is performed on the normalized matrix to obtain the second feature vector.

[0091] The standardized matrix X norm The data is directly input into the second convolutional neural network branch, which has a similar or identical structure to the first branch and is used to extract the spatial texture feature vector (i.e., the second feature vector). .

[0092] This step ensures that the model does not lose the intuitive morphological features in the normalized matrix due to an overemphasis on frequency domain information.

[0093] Step S4: Adaptive frequency-space feature decoupling and fusion.

[0094] To enable the model to automatically determine whether the current phase drift is severe and dynamically adjust the weights of the two branches, an adaptive gating fusion mechanism is used to fuse the first feature vector and the second feature vector.

[0095] Figure 3 This is a flowchart of an optional method for determining a target feature vector according to an embodiment of this application, such as... Figure 3 As shown, the steps for determining the target feature vector include:

[0096] Step S41, Feature concatenation, concatenating the high-dimensional frequency domain feature vector. Spatial texture feature vector By concatenating the channels, we obtain the concatenated feature vector.

[0097] Step S42: Input the concatenated feature vector into a fully connected layer (FC), and pass it through a sigmoid activation function to output a scalar weight. (i.e., initial weights) This represents the model's "degree of trust" in high-dimensional frequency domain feature vectors.

[0098] Step S43: Calculate the final diagnostic feature vector (i.e., the target feature vector). .

[0099]

[0100] Step S5: Classification and output of partial discharge defect types.

[0101] Step S51, the fused target feature vector Input the trained classifier (usually composed of fully connected layers and softmax layers).

[0102] Step S52: Output the probability distribution of various candidate partial discharge defect types (such as metal spikes, insulator air gaps, floating potentials, free particles, normal operation, etc.).

[0103] Step S53: Select the candidate partial discharge defect type with the highest probability as the final diagnosis result (i.e., the target partial discharge defect type).

[0104] The above optional implementation methods achieve at least the following effects:

[0105] (1) Phase drift interference was eliminated, and accurate partial discharge diagnosis with "zero synchronization" dependence was achieved. By utilizing the translation invariance principle of the amplitude spectrum of one-dimensional discrete Fourier transform, the phase drift problem was transformed into a mathematical identity transformation. This enabled the extraction of stable and consistent feature vectors even under harsh conditions where there was no power frequency synchronization signal or the phase drifted randomly at any angle (0°~360°). This reduced the hardware cost and operational risk of on-site detection and fundamentally ensured the high robustness of partial discharge diagnosis.

[0106] (2) Achieve feature decoupling between "morphological texture" and "frequency domain features" to balance the robustness and precision of recognition. By constructing a parallel dual-stream architecture of "frequency domain invariant + spatial domain texture", the frequency domain branch is responsible for locking periodic patterns (anti-drift), and the spatial domain branch is responsible for capturing local micro-textures (preserving accuracy), which solves the limitations of single-dimensional feature extraction. Even under severe phase drift, the frequency domain branch can provide strong classification support; while when the phase is relatively stable, the spatial domain branch can provide rich detailed information, thus maintaining a very high partial discharge diagnosis accuracy under various working conditions.

[0107] (3) It eliminates the dependence on data augmentation and has stronger generalization ability and physical interpretability. The introduction of deterministic mathematical transformation (DFT) instead of random sample augmentation significantly reduces the data threshold and computational consumption for model training. At the same time, this method based on physical transformation has clear mathematical meaning. Compared with the "black box" deep learning training, this alternative implementation method has theoretically 100% translational adaptability when facing unknown data, and its generalization ability is significantly improved.

[0108] (4) An adaptive gating fusion mechanism is introduced to improve the dynamic adaptability to complex field environments. The adaptive gating fusion mechanism can automatically calculate the weight ratio of the frequency domain and the spatial domain based on the input spliced ​​feature vector. When a severe phase drift is detected (spatial domain feature disorder), the weight of the high-dimensional frequency domain feature vector is automatically increased; otherwise, the two are used in a balanced manner. This dynamic adjustment mechanism further improves the ability to diagnose partial discharge types in complex electromagnetic environments.

[0109] 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, and 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.

[0110] This embodiment also provides a partial discharge defect type determination device for gas-insulated switchgear. This device is used to implement the above embodiments and preferred embodiments, and 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.

[0111] According to an embodiment of this application, an apparatus embodiment for implementing a method for determining the type of partial discharge defects in gas-insulated switchgear is also provided. Figure 4 This is a schematic diagram of an optional partial discharge defect type determination device for gas-insulated switchgear according to an embodiment of this application, as shown below. Figure 4 As shown, the partial discharge defect type determination device for the above-mentioned gas-insulated switchgear includes a data acquisition module 402, a first determination module 404, a second determination module 406, and a third determination module 408. The device will be described below.

[0112] Data acquisition module 402 is used to acquire the partial discharge pulse sequence of the target gas-insulated switchgear in the target cycle;

[0113] The first determining module 404, connected to the data acquisition module 402, is used to determine the phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence. The row of the phase-resolved partial discharge matrix represents the amplitude of the partial discharge pulse, and the column represents the phase of the partial discharge pulse. The phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulse of the target gas-insulated switchgear in the target period.

[0114] The second determining module 406, connected to the first determining module 404, is used to determine the target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix.

[0115] The third determining module 408, connected to the second determining module 406, is used to determine the type of target partial discharge defect of the target gas-insulated switchgear in the target period based on the target feature vector.

[0116] This application provides a device for determining the partial discharge defect type of a gas-insulated switchgear. By setting a data acquisition module 402, a first determination module 404, a second determination module 406, and a third determination module 408, the device aims to determine the phase-resolved partial discharge matrix based on the acquired partial discharge pulse sequence of the target gas-insulated switchgear, thereby obtaining the target partial discharge defect type of the target gas-insulated switchgear. This achieves the technical effect of improving the accuracy of the determination result of the target partial discharge defect type of the target gas-insulated switchgear, and solves the technical problem of inaccurate determination result of the partial discharge defect type of gas-insulated switchgear in related technologies.

[0117] 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.

[0118] It should be noted that the data acquisition module 402, the first determining module 404, the second determining module 406, and the third determining module 408 mentioned above correspond to steps S102 to S108 in the embodiments. The instances and application scenarios implemented by the above modules and their corresponding steps are the same, but they 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.

[0119] 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.

[0120] The partial discharge defect type determination device of the above-mentioned gas-insulated switchgear may also include a processor and a memory. The data acquisition module 402, the first determination module 404, the second determination module 406, the third determination module 408, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize the corresponding functions.

[0121] The processor contains a core that retrieves the corresponding program unit from memory. One or more cores may be configured. Memory may include non-persistent memory in computer-readable media, 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.

[0122] This application provides a non-volatile storage medium storing a program that, when executed by a processor, implements a method for determining the type of partial discharge defect in a gas-insulated switchgear.

[0123] This application provides an electronic device including a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it performs the following steps: acquiring a partial discharge pulse sequence of a target gas-insulated switchgear in a target period; determining a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulses, and the columns represent the phases of the partial discharge pulses, and the phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulses of the target gas-insulated switchgear in the target period; determining a target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix; and determining the target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector. The device described herein may be a server, PC, etc.

[0124] This application also provides a computer program product, which, when executed on a data processing device, is suitable for executing an initialization program having the following method steps: acquiring a partial discharge pulse sequence of a target gas-insulated switchgear in a target period; determining a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulses, and the columns represent the phases of the partial discharge pulses, and the phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulses of the target gas-insulated switchgear in the target period; determining a target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix; and determining a target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector.

[0125] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0126] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0127] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0128] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0129] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0130] 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 is an example of computer-readable media.

[0131] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can store information using any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase-change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, CD-ROM, digital versatile optical disc (DVD) or other optical storage, magnetic tape, magnetic disk storage or other magnetic storage devices, or any other non-transferable medium that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transient computer-readable media, such as modulated data signals and carrier waves.

[0132] It should also be noted that the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0133] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0134] The above are merely embodiments of this application and are not intended to limit the scope of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for determining the type of partial discharge defect in a gas-insulated switchgear, characterized in that, include: Obtain the partial discharge pulse sequence of the target gas-insulated switchgear during the target cycle; Based on the partial discharge pulse sequence, a phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period is determined, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulse, and the columns represent the phase of the partial discharge pulse. The phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulse of the target gas-insulated switchgear in the target period. Based on the phase-resolved partial discharge matrix, the target feature vector of the target gas-insulated switchgear in the target period is determined; Based on the target feature vector, the type of target partial discharge defect of the target gas-insulated switchgear in the target period is determined.

2. The method according to claim 1, characterized in that, The determination of the target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix includes: The phase-resolved partial discharge matrix is ​​normalized to obtain a standardized matrix; The standardized matrix is ​​reconstructed to obtain a reconstructed matrix, wherein each element in the reconstructed matrix is ​​a real number used to characterize the intensity of the partial discharge defect at the corresponding amplitude level; Feature extraction is performed on the reconstructed matrix to obtain the first feature vector of the target gas-insulated switchgear in the target period; Feature extraction is performed on the standardized matrix to obtain the second feature vector of the target gas-insulated switchgear in the target period; The target feature vector is determined based on the first feature vector and the second feature vector.

3. The method according to claim 2, characterized in that, The process of reconstructing the standardized matrix to obtain the reconstructed matrix includes: Perform a discrete Fourier transform on the normalized matrix to obtain a frequency domain matrix, wherein each element in the frequency domain matrix is ​​a complex number, used to characterize the amplitude and phase information of the harmonic components contained in the partial discharge pulse within the target period at the corresponding amplitude level; The reconstructed matrix is ​​obtained by extracting the magnitude of each element in the frequency domain matrix.

4. The method according to claim 2, characterized in that, Determining the target feature vector based on the first feature vector and the second feature vector includes: The first feature vector and the second feature vector are concatenated to obtain a concatenated feature vector. Based on the concatenated feature vector, the initial weight of the first feature vector is determined; The initial weights are corrected to obtain the target weights of the first feature vector; Based on the target weight, the first feature vector and the second feature vector are fused to obtain the target feature vector.

5. The method according to claim 4, characterized in that, The step of correcting the initial weights to obtain the target weights of the first feature vector includes: Based on the first feature vector, the frequency domain energy concentration is determined, wherein the frequency domain energy concentration is used to characterize the degree of aggregation of the power frequency fundamental wave and low harmonic components in the reconstruction matrix; Based on the second feature vector, the spatial texture disorder is determined, wherein the spatial texture disorder is used to quantify the average gray entropy of the normalized matrix, and the average gray entropy is used to describe the degree of local structural blurring of the normalized matrix caused by phase drift or noise interference. The initial weights are corrected based on the frequency domain energy concentration and the spatial domain texture disorder to obtain the target weights.

6. The method according to any one of claims 1 to 5, characterized in that, The determination of the target partial discharge defect type of the target gas-insulated switchgear in the target period based on the target feature vector includes: Based on the target feature vector, the probabilities corresponding to various candidate partial discharge defect types of the target gas-insulated switchgear are determined during the target period. The candidate partial discharge defect type with the highest probability among the probabilities corresponding to the various candidate partial discharge defect types is determined as the target partial discharge defect type.

7. A device for determining the type of partial discharge defect in a gas-insulated switchgear, characterized in that, include: The data acquisition module is used to acquire the partial discharge pulse sequence of the target gas-insulated switchgear during the target cycle; The first determining module is used to determine the phase-resolved partial discharge matrix of the target gas-insulated switchgear in the target period based on the partial discharge pulse sequence, wherein the rows of the phase-resolved partial discharge matrix represent the amplitude of the partial discharge pulse, the columns represent the phase of the partial discharge pulse, and the phase-resolved partial discharge matrix is ​​used to describe the distribution characteristics of the amplitude and phase of the partial discharge pulse of the target gas-insulated switchgear in the target period. The second determining module is used to determine the target feature vector of the target gas-insulated switchgear in the target period based on the phase-resolved partial discharge matrix. The third determining module is used to determine the type of target partial discharge defect of the target gas-insulated switchgear in the target period based on the target feature vector.

8. A non-volatile storage medium, characterized in that, The non-volatile storage medium stores multiple instructions adapted for loading and execution by a processor of the method for determining the partial discharge defect type of the gas-insulated switchgear according to any one of claims 1 to 6.

9. An electronic device, characterized in that, include: 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 method for determining the partial discharge defect type of a gas-insulated switchgear according to any one of claims 1 to 6.

10. A computer program product comprising computer instructions, characterized in that, When the computer instructions are executed by the processor, they implement the method for determining the partial discharge defect type of the gas-insulated switchgear as described in any one of claims 1 to 6.

Images