A transformer bushing arc fault early warning and operating state monitoring method and related device

Abstract

The embodiment of the application provides a transformer bushing arc fault early warning and operation state monitoring method and related equipment, and belongs to the technical field of intelligent monitoring of power equipment. The method comprises the following steps: acquiring operation image data and current voltage data of a target transformer bushing, inputting the related data into a trained early warning model for calculation, and obtaining a monitoring result of the operation state of the target transformer. Wherein, the method can evaluate the visual and electrical fault risks of the real-time data of the transformer bushing through the intelligent model, and further analyze the evaluation result according to a hierarchical weighted fusion strategy to obtain the monitoring result of the operation state of the transformer bushing, so as to realize the real-time monitoring of the operation state of the transformer bushing and effectively improve the monitoring sensitivity.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology for power equipment, and in particular to a method and related equipment for early warning of arc faults in transformer bushings and monitoring of their operating status. Background Technology

[0002] As a crucial link connecting the internal windings of a transformer to the external power grid, the transformer bushing continuously endures the combined effects of high voltage, high current, and complex mechanical stress during operation, making its working conditions extremely harsh. A failure in this component can easily trigger a deflagration of the transformer itself, leading to significant economic losses and power outages. Therefore, effective monitoring of its operating status is of paramount importance.

[0003] Among related technologies, the mainstream monitoring methods mainly include periodic offline testing, online electrical monitoring, and manual inspection and infrared thermography. Periodic offline testing requires a power outage and assesses the insulation status by measuring parameters such as dielectric loss factor and capacitance, but it cannot achieve real-time perception of the operating status. Online electrical monitoring usually uses end-screen sensors to collect leakage current, but this method is easily affected by ambient temperature and humidity, power grid harmonics, and electromagnetic interference, and there is a risk of false alarms and missed alarms. Methods relying on manual inspection or infrared imaging are not only inefficient, but also difficult to effectively capture transient events such as surface flashover or early microscopic defects in insulation materials, and have significant deficiencies in sensitivity.

[0004] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0005] The main objective of this application is to propose a method and related equipment for early warning of arc faults and monitoring of the operating status of transformer bushings. This method can assess the visual and electrical fault risks respectively through an intelligent model based on real-time data of transformer bushings, and further analyze the assessment results according to a hierarchical weighted fusion strategy to obtain the monitoring results of the operating status of transformer bushings, thereby realizing real-time monitoring of the operating status of transformer bushings and effectively improving monitoring sensitivity.

[0006] To achieve the above objectives, one aspect of this application proposes a method for early warning of transformer bushing arc faults and monitoring of operating status, the method comprising: Acquire operational image data and current and voltage data of the target transformer bushing; The running image data and the current and voltage data are input into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status; The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result using a hierarchical weighted fusion strategy to output a monitoring result of the operating status of the target transformer bushing.

[0007] In some embodiments, the process by which the visual fault analysis module outputs visual fault assessment results includes: Target recognition is performed on the operating image data of the target transformer bushing based on a lightweight neural network, and the region of interest data of the target transformer bushing is output. The region of interest data is first identified, and feature confidence is extracted based on the identification results to obtain static defect features; based on the static defect features, a weighted summation calculation and normalization are performed to obtain the static defect probability. The region of interest data is subjected to a second identification, and feature confidence is extracted and normalized based on the identification results to obtain dynamic anomaly features; the dynamic anomaly features include arc breakdown features and smoke features; the dynamic anomaly features are analyzed based on maximum value logic to obtain dynamic anomaly probabilities; Based on the analysis of the static defect probability and the dynamic anomaly probability, a visual fault assessment result is output.

[0008] In some embodiments, the process by which the electrical fault analysis module outputs electrical fault assessment results includes: The current and voltage data of the target transformer bushing are denoised using the wavelet threshold denoising method to obtain the denoised current and voltage data. Based on the wavelet transform algorithm, signal abrupt points in the denoised current and voltage data are identified to obtain the signal abrupt changes. High-frequency energy entropy data is obtained by calculating based on the wavelet packet decomposition algorithm and the denoised current and voltage data; the high-frequency energy entropy data is used to determine whether there is a discharge phenomenon in the transformer bushing. Based on the analysis of the signal abrupt changes and the high-frequency energy entropy data, an electrical fault assessment result is output.

[0009] In some embodiments, the process by which the comprehensive condition assessment module outputs the monitoring results of the target transformer bushing operating status includes: Based on the visual fault assessment results, analyze whether there are visual monitoring faults; based on the electrical fault assessment results, analyze whether there are electrical monitoring faults. If a visual monitoring fault or an electrical monitoring fault is detected, the output monitoring result will be an alarm. If there are no visual monitoring faults and no electrical monitoring faults, the fusion health index is obtained by weighting the visual fault assessment results and the electrical fault assessment results. The judgment is made based on the preset threshold and the fusion health index, and the monitoring result is output according to the judgment result.

[0010] In some embodiments, the step of judging based on a preset threshold and the fused health index, and outputting monitoring results based on the judgment result, includes: If the fusion health index is greater than the first preset threshold, the monitoring result is output as normal. If the fusion health index is less than or equal to the first preset threshold and greater than the second preset threshold, the monitoring result is output as abnormal. If the fusion health index is less than or equal to the second preset threshold, the output detection result is an alarm.

[0011] In some embodiments, the monitoring method further includes: Determine whether there are foreign objects based on the operating image data of the target transformer bushing; Analyze whether there is an electrical monitoring fault based on the electrical fault assessment results of the target transformer bushing; If there is a foreign object and there is no electrical monitoring fault, the output monitoring result will indicate the presence of external interference.

[0012] In some embodiments, the early warning model is trained in the following manner: Acquire a sample set of running image data, a sample set of current and voltage data, and a sample set of monitoring results; The running image data sample set and the current and voltage data sample set are input into the early warning pre-model for calculation to obtain the monitoring result training set; The monitoring results training set and the monitoring results sample set are analyzed, and the model parameters of the early warning model are adjusted according to the analysis results until the preset requirements are met. The trained early warning model is obtained based on the adjustment results.

[0013] To achieve the above objectives, another aspect of this application proposes a transformer bushing arc fault early warning and operation status monitoring system, the monitoring system including a data acquisition module and a monitoring module; wherein, The data acquisition module is used to acquire the operating image data and current and voltage data of the target transformer bushing; The monitoring module is used to input the operating image data and the current and voltage data into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status; The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result using a hierarchical weighted fusion strategy to output a monitoring result of the operating status of the target transformer bushing.

[0014] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the method described above.

[0015] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the methods described above.

[0016] The embodiments of this application include at least the following beneficial effects: This application provides a method, system, electronic device, and storage medium for early warning of arc faults and monitoring of the operating status of transformer bushings. This solution acquires the operating image data and current and voltage data of the target transformer bushing, analyzes the relevant data through a trained early warning model, assesses the visual fault risk and electrical fault risk of the target transformer bushing separately during the analysis, and further analyzes the two risk assessment results according to a hierarchical fusion strategy to output the monitoring results of the operating status of the target transformer bushing. On the one hand, unlike traditional monitoring methods that require power outages, the method of this application acquires transformer data in real time. The method generates relevant data on the bushing and outputs the monitoring results of its operating status, enabling real-time monitoring of the target transformer bushing with low requirements for monitoring environmental conditions. On the other hand, by separately assessing the visual and electrical fault risks of the target transformer bushing, the method can consider the importance of different types of fault risk assessments for monitoring the operating status of the transformer bushing, thereby improving the monitoring sensitivity of the transformer bushing's operating status. Furthermore, the method uses a hierarchical fusion strategy to comprehensively analyze the assessment results to obtain the final monitoring results, which can more accurately locate the fault risk of the transformer bushing and further improve the monitoring sensitivity of the transformer bushing's operating status. Attached Figure Description

[0017] Figure 1 This is a flowchart of a method for early warning of transformer bushing arc faults and monitoring of operating status provided in an embodiment of this application; Figure 2 This is a structural block diagram of an early warning model provided in an embodiment of this application; Figure 3 This is a structural block diagram of a transformer bushing arc fault early warning and operation status monitoring system provided in an embodiment of this application; Figure 4 This is a structural block diagram of an electronic device provided in an embodiment of this application. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0019] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0020] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0021] 1) RGB image: A digital image display method that synthesizes most of the colors in nature by superimposing the brightness values ​​of the three color channels: red (R), green (G), and blue (B).

[0022] 2) Lightweight Convolutional Neural Network (LCNN): A type of convolutional neural network that reduces the number of parameters and computational complexity through special structural designs such as depthwise separable convolution and model pruning, thereby maintaining high accuracy while being suitable for deployment on resource-constrained devices such as mobile devices.

[0023] 3) Wavelet packet decomposition (WPD): A further extension of wavelet transform, it not only decomposes the low-frequency part of the signal, but also iteratively decomposes the high-frequency part, thereby achieving more refined analysis and feature extraction of the entire frequency band of the signal.

[0024] 4) Dempster-Shafer Theory: A mathematical theory that describes and handles uncertainty by constructing trust functions and likelihood functions, and uses Dempster's synthesis rule to fuse information from multiple independent evidence sources for reasoning and decision-making.

[0025] 5) ResNet-50: A 50-layer deep convolutional neural network based on residual learning and constructed through shortcut connections. It can effectively solve the gradient vanishing problem in deep networks and is often used as a pre-trained backbone network for tasks such as image classification.

[0026] 6) Gaussian Mixture Model (GMM): A probabilistic model that uses a weighted combination of the probability density functions of multiple Gaussian distributions to describe the data distribution, and can effectively fit complex data distributions of arbitrary shapes.

[0027] 7) ROI (Region of Interest) refers to a specific region in an image that needs to be processed in detail, either through algorithms or manually, in image processing or computer vision tasks.

[0028] In related technologies, transformer bushings, as a crucial link connecting the internal windings of a transformer to the external power grid, continuously endure the combined effects of high voltage, high current, and complex mechanical stress during operation, making their working conditions extremely harsh. A failure in this component can easily trigger a deflagration of the transformer itself, leading to significant economic losses and power outages. Therefore, effective monitoring of its operating status is of paramount importance.

[0029] Currently, the mainstream methods for monitoring the operating status of transformer bushings have the following two shortcomings: (1) Insufficient real-time performance and sensitivity: Mainstream monitoring methods include periodic offline testing, online electrical monitoring, and manual inspection and infrared thermography. Periodic offline testing requires a power outage and assesses insulation status by measuring parameters such as dielectric loss factor and capacitance, but it cannot achieve real-time perception of operational status. Online electrical monitoring typically uses end-screen sensors to collect leakage current, but this method is susceptible to environmental temperature and humidity, power grid harmonics, and electromagnetic interference, posing a risk of false alarms and missed alarms. Methods relying on manual inspection or infrared imaging are not only inefficient but also struggle to effectively capture transient events such as surface flashovers or early microscopic defects in insulation materials, exhibiting significant deficiencies in real-time performance and sensitivity.

[0030] (2) Existing monitoring systems generally rely on a single sensing mode, which has significant limitations in terms of systematicness and reliability: While electrical parameter monitoring can effectively reflect internal insulation degradation and some fault characteristics, it struggles to identify physical defects such as mechanical damage and surface cracks caused by external stress. Visual monitoring, on the other hand, can directly capture visible features like deformation, displacement, and arc discharge, but lacks the ability to directly perceive latent internal defects such as insulation aging and localized overheating. This fragmentation of information dimensions leads to an incomplete understanding of the equipment's condition. Especially in extreme cases of sudden faults or rapid degradation, single monitoring systems, due to their limited sensing dimensions and insufficient information fusion, often fail to provide rapid and accurate fault warnings and diagnoses, thus failing to meet the engineering requirements for panoramic and intelligent condition assessment of transformer bushings.

[0031] In view of this, this application provides a method, system, electronic device, and storage medium for early warning and monitoring of arc faults in transformer bushings. This solution acquires operational image data and current and voltage data of the target transformer bushing, analyzes the relevant data using a trained early warning model, and assesses both visual and electrical fault risks of the target transformer bushing during the analysis. Then, it further analyzes the two risk assessment results based on a hierarchical fusion strategy to output the monitoring results of the target transformer bushing's operational status. On the one hand, unlike traditional monitoring methods that require power outages, this method can collect relevant data from the transformer bushing in real time and output its operational status monitoring results, enabling real-time monitoring of the target transformer bushing with lower requirements for monitoring environmental conditions. On the other hand, by separately assessing the visual and electrical fault risks of the target transformer bushing, this method can consider the importance of different types of fault risk assessments for monitoring the transformer bushing's operational status, thereby improving the monitoring sensitivity of the transformer bushing's operational status. Furthermore, this method uses a hierarchical fusion strategy to comprehensively analyze the assessment results to obtain the final monitoring result, which can more accurately locate the fault risk of the transformer bushing and further improve the monitoring sensitivity of the transformer bushing's operational status.

[0032] The transformer bushing operation status monitoring method provided in this application relates to the field of intelligent monitoring technology for power equipment. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle-mounted terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the transformer bushing operation status monitoring method, but is not limited to the above forms.

[0033] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0034] Figure 1 This is an optional flowchart of a transformer bushing arc fault early warning and operation status monitoring method provided in this application embodiment. The structural block diagram of the early warning model corresponding to this method is as follows: Figure 2 As shown; Figure 1 The method may include, but is not limited to, steps S101 to S102.

[0035] Step S101: Obtain the operating image data and current and voltage data of the target transformer bushing.

[0036] Obtain relevant data on the target transformer bushing as the data foundation.

[0037] Step S102: Input the running image data and current and voltage data into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status.

[0038] The image data and current and voltage data can be obtained in real time or in historical data. By inputting the relevant data into the model, the monitoring results of the operating status within the corresponding monitoring time period can be obtained. In this respect, unlike traditional monitoring methods that require power outages, the method of this application can collect relevant data of transformer bushings in real time and output the monitoring results of their operating status. It can realize real-time monitoring of the target transformer bushings and has low requirements for monitoring environmental conditions.

[0039] The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result based on a hierarchical weighted fusion strategy and outputs the monitoring result of the operating status of the target transformer bushing.

[0040] First, the visual fault analysis module assesses the visual fault risks of the target transformer bushing, such as cracks, oil stains, electric arcs, and smoke. Then, the electrical fault analysis module assesses the electrical faults of the target transformer bushing, such as sudden current or voltage changes. Finally, the comprehensive condition assessment module analyzes the fault risk assessment results based on a hierarchical weighted fusion strategy, and outputs the monitoring results of the target transformer bushing.

[0041] In some embodiments, the acquisition of operational image data and current and voltage data of the target transformer bushing includes basic data acquisition and data preprocessing: (1) Basic data collection: The operating image data is acquired by the image acquisition system: an industrial-grade dual-spectrum camera (1920×1080 resolution, 60fps) is used to capture RGB images of the transformer bushings during their operation. The camera has fog-penetrating and strong light suppression functions to adapt to the outdoor substation environment.

[0042] Current and voltage data are acquired by a current and voltage oscilloscope: Current sampling: A current sensor is installed on the bushing grounding lead to collect power frequency leakage current and high frequency partial discharge pulses; Voltage sampling: A voltage sensor is used to acquire a synchronous voltage signal as a phase reference.

[0043] (2) Data preprocessing: The acquired experimental data were preprocessed using methods such as data cleaning, data denoising, and format standardization to reduce the impact of noise in the data on subsequent research and prediction.

[0044] In some embodiments, the process by which the visual fault analysis module outputs visual fault assessment results includes, but is not limited to, steps S201 to S204.

[0045] Step S201: Based on a lightweight neural network, target recognition is performed on the operating image data of the target transformer bushing, and the region of interest data of the target transformer bushing is output.

[0046] By processing the operating image data of the target transformer bushing through a lightweight neural network, more targeted region of interest data is output as the data basis for subsequent processing.

[0047] Optionally, a lightweight YOLOv8-Nano network can be used to perform inference on the input sleeve video frames, outputting the sleeve's bounding box. This bounding box is used to extract the transformer sleeve portion from the image, generating the sleeve's Region of Interest (ROI). The generation of ROI data aims to reduce interference from the background and other background information, forcing the model to focus on the sleeve region within the video frame. By cropping the sleeve ROI from the video frame, the influence of the background during recognition can be effectively reduced, lowering the false alarm rate caused by background interference, while simultaneously reducing the computational requirements of the device and improving processing speed.

[0048] Furthermore, visual fault analysis includes both static defects and dynamic anomalies. The method in this application adopts a strategy that combines hierarchical weighting and static and dynamic faults. The processed region of interest data is input into two parallel visual processing channels for processing, thereby obtaining the probability of static defects and the probability of dynamic anomalies. The two probabilities are then weighted and processed to finally output the visual fault assessment result.

[0049] Step S202: Perform a first identification on the region of interest data and extract feature confidence based on the identification results to obtain static defect features; perform weighted summation and normalization based on the static defect features to obtain the static defect probability.

[0050] Optionally, the region of interest data is identified through a static defect branch channel to identify static defects (external defects), such as cracks and oil stains. Feature confidence is then extracted from the identification results to obtain static defect features, such as crack features. and oil stain characteristics wait.

[0051] Furthermore, for static defects such as cracks and oil stains, since these types of defects do not disappear instantly, after extracting feature confidence scores, a weighted sum is used and normalized using a sigmoid function to obtain the static defect probability: (1) in This represents the static defect probability. , Represents the coefficient. + =1, given the characteristics of static defects, The values ​​are 0.5 and 0.5 respectively.

[0052] Step S203: Perform a second identification on the region of interest data, and extract and normalize the feature confidence based on the identification results to obtain dynamic anomaly features; the dynamic anomaly features include arc breakdown features. and smoke characteristics The dynamic anomaly characteristics are analyzed based on the maximum value logic to obtain the dynamic anomaly probability.

[0053] Optionally, dynamic anomalies can be identified in the region of interest data through dynamic anomaly molecular channels, such as drastic brightness changes caused by arc breakdown and dynamic anomalies like smoke. The identification results are then further processed to extract feature confidence, ultimately yielding dynamic anomaly features, such as arc breakdown features. and smoke characteristics .

[0054] Furthermore, for dynamic anomalies such as arc breakdown characteristics and smoke characteristics, since the appearance of these characteristics often indicates a possible fault in the transformer bushing, a veto strategy is adopted for this type of anomaly. That is, as long as any anomaly is detected, an alarm is issued directly. Therefore, maximum value logic is used. max( , (2) in This represents the probability of dynamic anomalies. An alarm is issued as soon as the model detects either of these two anomalies.

[0055] Step S204: Analyze the static defect probability and dynamic anomaly probability, and output the visual fault assessment result.

[0056] By combining the dual-path analysis of static defects and dynamic anomalies, and then comprehensively analyzing the results of the dual-path analysis, a more accurate visual fault assessment result is output.

[0057] Alternatively, the comprehensive analysis process can be achieved by the following equation (3): ( , (3) in, This represents the probability of visual failure. Since dynamic failures are more urgent than static failures, their probability should be assigned a higher priority; for example, if... If the probability exceeds 0.6, it is judged as a dynamic fault such as arc breakdown or smoke generation, and the visual fault assessment result is output as a warning.

[0058] To address this, the present invention employs a dual-path analysis mechanism that combines static defects and dynamic anomalies. The static defect branch utilizes a deep convolutional network to extract minute cracks and oil stains on the bushing surface, ensuring a high detection rate for insulation defects. The dynamic flow branch, based on bushing brightness variation detection and a Gaussian mixture model, specifically monitors the highly dynamic characteristics of arc and smoke diffusion.

[0059] In some embodiments, the process by which the electrical fault analysis module outputs electrical fault assessment results includes, but is not limited to, steps S301 to S304.

[0060] Step S301: Denoise the current and voltage data of the target transformer bushing based on the wavelet threshold denoising method to obtain the denoised current and voltage data.

[0061] Wavelet thresholding denoising method is used to denoise the acquired current signal and voltage signal Denoising is performed to remove noise from the signal in order to reduce interference with data analysis.

[0062] Step S302: Based on the wavelet transform algorithm, identify signal abrupt change points in the denoised current and voltage data to obtain the signal abrupt change situation.

[0063] Detection is performed using wavelet transform. The signal is decomposed into wavelet components of different scales and frequencies using wavelet transform, and the detail coefficients at different scales are analyzed to identify points of abrupt changes in the detail coefficients. An appropriate threshold is selected based on the signal's noise level and the significance of the abrupt change. The system automatically identifies sudden, sharp changes in the signal that are significantly inconsistent with the normal pattern or baseline, analyzes the energy or amplitude of the wavelet coefficients, accurately locates the moment of the abrupt change, and outputs the signal abrupt change information.

[0064] Step S303: Calculate high-frequency energy entropy data based on wavelet packet decomposition algorithm and denoised current and voltage data; the high-frequency energy entropy data is used to determine whether there is a discharge phenomenon in the transformer bushing.

[0065] Unlike FFT, which provides a general overview of the entire spectrum, wavelet packet decomposition breaks down the current waveform into sub-signals of different frequency bands. Leveraging this characteristic, wavelet packet decomposition divides the current signal into eight frequency bands, calculating the energy proportion of high-frequency nodes, i.e., high-frequency energy entropy data. When internal discharge occurs, the energy of the high-frequency components of the energy entropy increases significantly; therefore, calculating the energy entropy of the high-frequency sub-bands can determine whether a discharge has occurred. During normal operation, the signal is orderly with low entropy; once a weak discharge occurs, the signal becomes chaotic, and the entropy value of the high-frequency bands spikes instantly. By analyzing the energy entropy value, weak early partial discharge signals can be accurately identified even under the mask of strong power frequency current, enabling fault monitoring of transformer bushings.

[0066] Step S301: Analyze the signal abrupt changes and high-frequency energy entropy data to output electrical fault assessment results.

[0067] By combining signal abrupt changes with high-frequency energy entropy data for analysis, it is possible to ultimately determine whether an arc discharge phenomenon has occurred and output electrical fault assessment results.

[0068] Specifically, in actual operation, sudden current changes may not necessarily result in arc discharge. Therefore, if there is only a sudden current change but no discharge phenomenon is detected from the high-frequency energy entropy data, the electrical fault assessment result is normal operation, and the electrical fault probability is [not specified]. If a signal mutation is detected simultaneously from relevant data and a discharge phenomenon is detected in high-frequency energy entropy data, the output electrical fault assessment result is an alarm, and the electrical fault probability is... 1. Issue an alarm. It's important to note that since the presence or absence of discharge is a crucial factor in ensuring normal operation, any detected discharge will trigger an alarm in the electrical fault assessment, indicating an electrical fault probability. 1.

[0069] In response, when analyzing electrical signals such as current waveforms, the method of this invention is not limited to simple threshold determination of the effective value of the current. Instead, it introduces a joint time-frequency domain analysis mechanism. Wavelet packet decomposition is used to perform multi-level decomposition of the acquired current. The system extracts weak partial discharge pulse signals from the high-amplitude power frequency fundamental wave and calculates the energy entropy of the high-frequency sub-band as a key indicator for fault early warning, identifying whether a discharge phenomenon has occurred. The operating status of the transformer bushing is monitored through electrical information.

[0070] In some embodiments, the process of the comprehensive status assessment module outputting the monitoring results of the target transformer bushing operating status includes, but is not limited to, steps S401 to S404.

[0071] Step S401: Analyze whether there is a visual monitoring fault based on the visual fault assessment results and analyze whether there is an electrical monitoring fault based on the electrical fault assessment results.

[0072] First, analyze the results of the visual fault assessment and the electrical fault assessment to determine whether there is a visual monitoring fault or an electrical monitoring fault. If the visual fault assessment result shows that an electric arc or smoke is detected, it is determined that there is a visual monitoring fault. If the electrical fault assessment result shows that a sudden change in current or voltage and a discharge phenomenon are detected, it is determined that there is an electrical monitoring fault.

[0073] The method in this application outputs monitoring results based on a hierarchical weighted fusion strategy, defining the state set of the target transformer bushing. ;in, This is the normal state. This is an abnormal state. The system is in an alarm state. Rules including, but not limited to, Rule 1 (step S402) and Rule 2 (steps S403 to S404) are set: Step S402: If a visual monitoring fault or an electrical monitoring fault exists, the monitoring result is output as an alarm.

[0074] Rule 1: Explosion / Penetration Protection: If a visual or electrical monitoring fault is detected, the monitoring result will be output as an alarm. This rule has the highest priority in the entire monitoring output rules to ensure a fast response.

[0075] Step S403: If there is no visual monitoring fault and no electrical monitoring fault, a weighted calculation is performed based on the visual fault assessment results and the electrical fault assessment results to obtain the fusion health index.

[0076] Step S404: Make a judgment based on the preset threshold and the integrated health index, and output the monitoring result according to the judgment result.

[0077] Rule 2: Comprehensive assessment of latent faults (based on formula (4) to calculate the fusion health index) ): (4) in, and The weighting coefficient is determined based on the characteristics of the transformer bushing.

[0078] Optionally, it can be taken and In cases where the weight of electrical information is slightly greater than that of visual information, the internal current, voltage, and other electrical information become more crucial for determining the working status of the bushing.

[0079] Furthermore, based on preset thresholds and integrated health indices, a judgment is made, and monitoring results are output according to the judgment results, including but not limited to the following judgment steps: If the integrated health index is greater than the first preset threshold, the output monitoring result is normal.

[0080] If the integrated health index is less than or equal to the first preset threshold and greater than the second preset threshold, the output monitoring result is abnormal.

[0081] If the fusion health index is less than or equal to the second preset threshold, the output detection result is an alarm.

[0082] Optionally, the determination steps can be set as follows: like The output state is ; like The output state is (Abnormal state), and output early warning information that requires inspection; like The output state is The output state is (High-risk status, alarm), and output a warning message that a power outage is required for maintenance.

[0083] In some embodiments, the monitoring method further includes steps S601 to S603: Step S601: Determine whether there are foreign objects based on the operating image data of the target transformer bushing; Step S602: Analyze whether there is an electrical monitoring fault based on the electrical fault assessment results of the target transformer bushing; Step S603: If there is a foreign object and no electrical monitoring fault, the monitoring result is output as "external interference exists".

[0084] In response, the monitoring rules of this application's monitoring method also include Rule 3 (anti-interference logic): If the electrical signal is operating normally, but the visual image shows a foreign object passing by, the output state is... It is operating normally. Under actual operating conditions, it is determined to be external environmental interference, and the electrical false alarm is automatically blocked.

[0085] In some embodiments, the early warning model is trained in the following manner: Step S701: Obtain the running image data sample set, current and voltage data sample set, and monitoring result sample set.

[0086] The aim of acquiring multi-source heterogeneous data, including runtime image samples, current and voltage data samples, and monitoring result samples, is to construct a mapping between data characteristics and risks. ,in This represents the mapping relationship used in model construction. Represents the learnable parameters of the model. and These represent the RGB image data and current / voltage waveform data of the target transformer bushing, respectively.

[0087] Step S702: Input the running image data sample set and the current and voltage data sample set into the early warning pre-model for calculation to obtain the monitoring result training set.

[0088] Step S703: Analyze the monitoring results training set and the monitoring results sample set, adjust the model parameters of the early warning model according to the analysis results until the preset requirements are met, and obtain the trained early warning model according to the adjustment results.

[0089] Using optimization algorithms such as SGD and Adam, and based on the analysis between the training set and the sample set of monitoring results, the model parameters (such as the learning rate) are dynamically adjusted through an exponential decay strategy to train the constructed machine learning model, obtain the optimal parameters of the model, and finally obtain an early warning model that can predict the operating status and potential risks of transformer bushings.

[0090] In summary, the embodiments of this method include, but are not limited to, the following beneficial effects: (1) Improve the real-time monitoring of transformer bushings: Unlike traditional monitoring methods that require power outages, the method in this application can achieve real-time monitoring of the target transformer bushing by collecting relevant data of the transformer bushing in real time and outputting the monitoring results of its operating status.

[0091] (2) Reduce the requirements for monitoring environmental conditions: Traditional real-time monitoring methods are susceptible to environmental temperature and humidity, power grid harmonics, and electromagnetic interference, which pose risks of false alarms and missed alarms. In contrast, the method in this application only requires the collection of relevant data, which can be input into the early warning model to output the corresponding monitoring results, thus reducing the requirements for the monitoring environment.

[0092] (3) Improve monitoring sensitivity: On the one hand, the method of this application, by separately assessing the visual and electrical fault risks of the target transformer bushing, can take into account the importance of different types of fault risk assessments for monitoring the operating status of the transformer bushing, thereby improving the monitoring sensitivity of the operating status of the transformer bushing; furthermore, the method of this application, based on a hierarchical fusion strategy, comprehensively analyzes the assessment results to obtain the final monitoring results, which can more accurately locate the fault risk of the transformer bushing, further improving the monitoring sensitivity of the operating status of the transformer bushing.

[0093] Please see Figure 3 This application also provides a transformer bushing arc fault early warning and operation status monitoring system, which can implement the above-mentioned method. The monitoring system includes a data acquisition unit and a monitoring unit; wherein... The data acquisition unit is used to acquire the operating image data and current and voltage data of the target transformer bushing; The monitoring unit is used to input the operating image data and current and voltage data into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status; The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result based on a hierarchical weighted fusion strategy and outputs the monitoring result of the operating status of the target transformer bushing.

[0094] It is understood that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0095] Please see Figure 4 This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0096] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0097] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0098] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0099] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0100] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0101] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0102] This application provides a method, system, electronic device, and storage medium for early warning of arc faults and monitoring of the operating status of transformer bushings. The solution acquires operating image data and current and voltage data of the target transformer bushing, analyzes the relevant data using a trained early warning model, and assesses both visual and electrical fault risks of the target transformer bushing during the analysis. Then, it further analyzes the two risk assessment results according to a hierarchical fusion strategy to output the monitoring results of the target transformer bushing's operating status. On the one hand, unlike traditional monitoring methods that require power outages, this method acquires relevant data of the transformer bushing in real time and outputs its operating status monitoring results, enabling real-time monitoring of the target transformer bushing with lower requirements for monitoring environmental conditions. On the other hand, by separately assessing the visual and electrical fault risks of the target transformer bushing, this method can consider the importance of different types of fault risk assessments for monitoring the transformer bushing's operating status, thereby improving the monitoring sensitivity of the transformer bushing's operating status. Furthermore, this method uses a hierarchical fusion strategy to comprehensively analyze the assessment results to obtain the final monitoring result, which can more accurately locate the fault risk of the transformer bushing and further improve the monitoring sensitivity of the transformer bushing's operating status.

[0103] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of this application, and do not constitute a limitation on the technical solutions provided in this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in this application are also applicable to similar technical problems.

[0104] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0105] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0106] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0107] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification 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.

[0108] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) of a, b, or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0109] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units 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; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0110] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0111] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0112] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, 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 storage medium and includes multiple 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 application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0113] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for early warning of arc faults in transformer bushings and monitoring of their operating status, characterized in that, The method includes the following steps: Acquire operational image data and current and voltage data of the target transformer bushing; The running image data and the current and voltage data are input into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status; The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result using a hierarchical weighted fusion strategy to output a monitoring result of the operating status of the target transformer bushing.

2. The monitoring method according to claim 1, characterized in that, The process by which the visual fault analysis module outputs visual fault assessment results includes: Target recognition is performed on the operating image data of the target transformer bushing based on a lightweight neural network, and the region of interest data of the target transformer bushing is output. The region of interest data is first identified, and feature confidence is extracted based on the identification results to obtain static defect features; based on the static defect features, a weighted summation calculation and normalization are performed to obtain the static defect probability. The region of interest data is subjected to a second identification, and feature confidence is extracted and normalized based on the identification results to obtain dynamic anomaly features; the dynamic anomaly features include arc breakdown features and smoke features; the dynamic anomaly features are analyzed based on maximum value logic to obtain dynamic anomaly probabilities; Based on the analysis of the static defect probability and the dynamic anomaly probability, a visual fault assessment result is output.

3. The monitoring method according to claim 1, characterized in that, The process by which the electrical fault analysis module outputs electrical fault assessment results includes: The current and voltage data of the target transformer bushing are denoised using the wavelet threshold denoising method to obtain the denoised current and voltage data. Based on the wavelet transform algorithm, signal abrupt points in the denoised current and voltage data are identified to obtain the signal abrupt changes. High-frequency energy entropy data is obtained by calculating based on the wavelet packet decomposition algorithm and the denoised current and voltage data; the high-frequency energy entropy data is used to determine whether there is a discharge phenomenon in the transformer bushing. Based on the analysis of the signal abrupt changes and the high-frequency energy entropy data, an electrical fault assessment result is output.

4. The monitoring method according to claim 1, characterized in that, The process by which the comprehensive condition assessment module outputs the monitoring results of the target transformer bushing operating status includes: Based on the visual fault assessment results, analyze whether there are visual monitoring faults; based on the electrical fault assessment results, analyze whether there are electrical monitoring faults. If a visual monitoring fault or an electrical monitoring fault is detected, the output monitoring result will be an alarm. If there are no visual monitoring faults and no electrical monitoring faults, the fusion health index is obtained by weighting the visual fault assessment results and the electrical fault assessment results. The judgment is made based on the preset threshold and the fusion health index, and the monitoring result is output according to the judgment result.

5. The monitoring method according to claim 4, characterized in that, The judgment based on a preset threshold and the integrated health index, and the output of monitoring results based on the judgment result, include: If the fusion health index is greater than the first preset threshold, the monitoring result is output as normal. If the fusion health index is less than or equal to the first preset threshold and greater than the second preset threshold, the monitoring result is output as abnormal. If the fusion health index is less than or equal to the second preset threshold, the output detection result is an alarm.

6. The monitoring method according to claim 1, characterized in that, The monitoring method also includes: Determine whether there are foreign objects based on the operating image data of the target transformer bushing; Analyze whether there is an electrical monitoring fault based on the electrical fault assessment results of the target transformer bushing; If there is a foreign object and there is no electrical monitoring fault, the output monitoring result will indicate the presence of external interference.

7. The monitoring method according to claim 1, characterized in that, The early warning model is trained in the following way: Acquire a sample set of running image data, a sample set of current and voltage data, and a sample set of monitoring results; The running image data sample set and the current and voltage data sample set are input into the early warning pre-model for calculation to obtain the monitoring result training set; The monitoring results training set and the monitoring results sample set are analyzed, and the model parameters of the early warning model are adjusted according to the analysis results until the preset requirements are met. The trained early warning model is obtained based on the adjustment results.

8. A transformer bushing arc fault early warning and operation status monitoring system, characterized in that, The monitoring system includes a data acquisition unit and a monitoring unit; wherein... The data acquisition unit is used to acquire the operating image data and current and voltage data of the target transformer bushing; The monitoring unit is used to input the operating image data and the current and voltage data into the trained early warning model for calculation to obtain the monitoring results of the target transformer bushing operating status; The early warning model includes a visual fault analysis module, an electrical fault analysis module, and a comprehensive status assessment module. The visual fault analysis module extracts features from the operating image data of the target transformer bushing and analyzes the results to output a visual fault assessment result. The electrical fault analysis module extracts features from the current and voltage data of the target transformer bushing and analyzes the results to output an electrical fault assessment result. The comprehensive status assessment module analyzes the visual fault assessment result and the electrical fault assessment result using a hierarchical weighted fusion strategy to output a monitoring result of the operating status of the target transformer bushing.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images