Transformer partial discharge monitoring method based on ultra-high frequency technology

Abstract

The present invention provides a transformer partial discharge monitoring method based on ultra-high frequency technology, which relates to the field of partial discharge monitoring technology, including: presetting a monitoring point array, deploying an ultra-high frequency sensor group for dynamic discharge monitoring, and obtaining an ultra-high frequency signal timing array; obtaining a multi-dimensional environmental signal timing array, performing filtering compensation, and obtaining a partial discharge signal timing array; performing distributed signal analysis, and obtaining a partial discharge characteristic parameter array; constructing a partial discharge characteristic tree, performing correlation matching to obtain a partial discharge cause characteristic parameter array; retrieving a partial discharge operation and maintenance strategy library; traversing matching to obtain a target partial discharge operation and maintenance strategy set, and performing partial discharge early warning operation and maintenance processing. The present invention solves the technical problem that the existing technology for transformer partial discharge monitoring relies more on experience judgment and manual analysis, lacks systematic diagnosis and automated processing decision support, resulting in low accuracy and efficiency of partial discharge processing.

Description

Technical Field

[0001] The invention relates to the technical field of partial discharge monitoring, and in particular to a transformer partial discharge monitoring method based on ultra-high frequency technology. Background Art

[0002] Transformers are vital equipment in power systems, and their reliable operation directly affects the stability and safety of power grids. However, during long-term operation, the insulation system of transformers may cause partial discharge due to various factors, such as aging, overload, and environmental changes. Partial discharge is one of the early manifestations of transformer insulation failure. If it is not discovered and handled in time, it may lead to further degradation of insulation and even cause serious electrical failures and equipment damage. Traditional partial discharge monitoring technologies, such as ultrasonic monitoring and electromagnetic wave monitoring, are greatly affected by environmental noise, electromagnetic interference and other factors when capturing partial discharge signals, and are prone to signal distortion or masking, resulting in insufficient monitoring accuracy and sensitivity. In addition, existing technologies for transformer partial discharge monitoring rely more on experience judgment and manual analysis, lack of systematic diagnosis and automated processing decision support, resulting in low accuracy and efficiency of partial discharge processing. Summary of the invention

[0003] The present application provides a transformer partial discharge monitoring method based on ultra-high frequency technology, aiming to solve the technical problem that the existing technology for transformer partial discharge monitoring relies heavily on experience judgment and manual analysis, lacks systematic diagnosis and automated processing decision support, resulting in low accuracy and efficiency of partial discharge processing.

[0004] The present application discloses a transformer partial discharge monitoring method based on ultra-high frequency technology, the method comprising: interactively obtaining a potential partial discharge point set of a target transformer, presetting a monitoring point array according to the potential partial discharge point set, deploying an ultra-high frequency sensor group based on the monitoring point array, performing dynamic discharge monitoring on the target transformer, and obtaining an ultra-high frequency signal timing array; deploying a multi-dimensional sensor group based on the monitoring point array, performing multi-dimensional dynamic environmental monitoring on the target transformer, and obtaining a multi-dimensional environmental signal timing array; filtering and compensating the ultra-high frequency signal timing array based on the multi-dimensional environmental signal timing array to obtain a partial discharge signal timing array; constructing a monitoring point data processing module set according to the monitoring point array, performing distributed signal analysis on the partial discharge signal timing array through the monitoring point data processing module set, and obtaining A partial discharge characteristic parameter array is constructed; a partial discharge characteristic tree is constructed, and an association matching is performed based on the partial discharge characteristic parameter array and the partial discharge characteristic tree to obtain a partial discharge cause characteristic parameter array; a partial discharge operation and maintenance instruction is triggered, and a partial discharge operation and maintenance strategy library is retrieved based on the partial discharge operation and maintenance instruction, wherein the partial discharge operation and maintenance strategy library is constructed based on historical partial discharge operation and maintenance records, and each record in the historical partial discharge operation and maintenance records includes historical partial discharge characteristic information, historical partial discharge cause characteristic information, and historical partial discharge defect operation and maintenance information; according to the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array, matching is traversed in the partial discharge operation and maintenance strategy library to obtain a target partial discharge operation and maintenance strategy set, and partial discharge early warning operation and maintenance processing is performed on the target transformer based on the target partial discharge operation and maintenance strategy set.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] Determine the potential local discharge point set of the target transformer and preset the monitoring point array to ensure that all areas where local discharge may occur are monitored. By deploying an ultra-high frequency sensor group for dynamic discharge monitoring, an ultra-high frequency signal timing array can be obtained, thereby accurately capturing the dynamic changes of the local discharge signal; by deploying a multi-dimensional sensor group for multi-dimensional dynamic environmental monitoring, a multi-dimensional environmental signal timing array is obtained, and based on these environmental signals, the ultra-high frequency signal timing array is filtered and compensated, effectively filtering out environmental noise and interference factors, improving the accuracy and credibility of the local discharge signal, and ensuring the high precision of the monitoring results; by constructing a set of monitoring point data processing modules, a distributed signal analysis is performed on the local discharge signal timing array to obtain a local discharge characteristic parameter array. The distributed processing method improves the speed and efficiency of data analysis, making it possible to analyze a large amount of monitoring data in real time, thereby quickly identifying local discharge characteristics; based on local discharge characteristics The parameter array constructs a local discharge feature tree and performs association matching to obtain a local discharge cause feature parameter array. The construction and matching process of this feature tree can systematically represent the relationship between local discharge features and causes, which helps to deeply understand the causes and development processes of local discharge and support more accurate fault diagnosis. By triggering local discharge operation and maintenance instructions and calling the operation and maintenance strategy library built based on historical local discharge operation and maintenance records, it can traverse and match in the strategy library according to the current local discharge feature parameters and cause feature parameters, and automatically generate a target local discharge operation and maintenance strategy set. This process realizes the automation and intelligence of operation and maintenance decision-making, making fault handling more efficient and targeted. Based on the target local discharge operation and maintenance strategy set, the target transformer is subjected to local discharge early warning operation and maintenance. This early warning mechanism can detect and handle potential local discharge faults at an early stage, effectively prevent the occurrence of major faults, and improve the operating safety and reliability of the equipment.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A schematic flow chart of a transformer partial discharge monitoring method based on ultra-high frequency technology is provided for an embodiment of the present application;

[0009] Figure 2 A schematic diagram of the process of obtaining operation and maintenance intervention priority in a transformer partial discharge monitoring method based on ultra-high frequency technology is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0010] The embodiments of the present application provide a transformer partial discharge monitoring method based on ultra-high frequency technology, thereby solving the technical problem that the existing technology for transformer partial discharge monitoring relies heavily on experience judgment and manual analysis, lacks systematic diagnosis and automated processing decision support, and leads to low accuracy and efficiency of partial discharge processing.

[0011] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0012] like Figure 1 As shown, the embodiment of the present application provides a transformer partial discharge monitoring method based on ultra-high frequency technology, the method comprising:

[0013] Interactively obtain a potential local discharge point set of the target transformer, preset a monitoring point array according to the potential local discharge point set, deploy an ultra-high frequency sensor group based on the monitoring point array, perform dynamic discharge monitoring on the target transformer, and obtain an ultra-high frequency signal timing array.

[0014] Through historical data analysis, expert evaluation and other interactive means, the potential partial discharge point set of the target transformer is obtained. The potential partial discharge point refers to the possible location where the partial discharge phenomenon exists in the transformer. Partial discharge is a local insulation damage phenomenon caused by insulation material defects or other reasons in electrical equipment. Identifying potential discharge points helps to prevent transformer insulation failures in advance. According to the potential partial discharge point set, the monitoring point array is preset. Each monitoring point represents the specific installation location of a sensor. The array needs to cover all potential partial discharge points as much as possible, while avoiding redundant coverage and blind spots of signals.

[0015] At the preset monitoring points, UHF sensor groups are deployed to capture partial discharge signals inside the transformer. UHF sensors are used to detect UHF electromagnetic wave signals generated by partial discharge. Since partial discharge generates high-frequency electromagnetic waves inside the transformer, UHF sensors can detect these signals with high sensitivity. After installation, the sensor begins to conduct continuous dynamic discharge monitoring of the target transformer, capturing the UHF signals inside the transformer in real time to form signal timing data. The timing array contains the signal timing data at each monitoring point.

[0016] A multi-dimensional sensor group is arranged based on the monitoring point array, multi-dimensional dynamic environmental monitoring is performed on the target transformer, a multi-dimensional environmental signal timing array is obtained, and filtering compensation is performed on the ultra-high frequency signal timing array based on the multi-dimensional environmental signal timing array to obtain a partial discharge signal timing array.

[0017] The multidimensional sensor group is a sensor group composed of various types of sensors. The corresponding sensor type can be selected according to the environmental factors that may affect the partial discharge signal. For example, temperature sensors, humidity sensors, vibration sensors, electric field and magnetic field sensors, etc. can simultaneously monitor multiple physical quantities around the transformer. The multidimensional sensor group is arranged based on the monitoring point array, so that the arrangement positions of these sensors can be consistent with the ultra-high frequency sensors, ensuring the monitoring of multi-dimensional environmental signals at the same position.

[0018] After the installation is completed, the multi-dimensional sensor group begins to conduct continuous dynamic environmental monitoring of the target transformer, collecting various environmental signals such as temperature, humidity, vibration, electric field, magnetic field, etc. in real time to form a multi-dimensional environmental signal time series array.

[0019] Since the partial discharge signal will be affected by the surrounding environment, such as temperature changes, vibration interference, etc., the directly acquired ultra-high frequency signal often contains noise and interference components. The purpose of filtering compensation is to extract pure partial discharge signals by removing or weakening these interferences. Specifically, the correlation between the multi-dimensional environmental signal and the ultra-high frequency signal is analyzed, and the interference mode of environmental factors on the partial discharge signal is identified. For example, the specific impact of factors such as temperature changes and vibrations on the ultra-high frequency signal is determined through correlation analysis. Filtering algorithms such as Kalman filtering, adaptive filtering, and wavelet denoising are applied to compensate the ultra-high frequency signal based on the multi-dimensional environmental signal timing array to remove environmental noise and interference signals. The ultra-high frequency signal after filtering compensation is the timing array of the partial discharge signal after the interference is removed. After processing, these signals can more accurately reflect the partial discharge activity inside the transformer, thereby providing a reliable data basis for subsequent fault diagnosis and operation and maintenance decisions.

[0020] A monitoring point data processing module set is constructed according to the monitoring point array, and a distributed signal analysis is performed on the partial discharge signal timing array through the monitoring point data processing module set to obtain a partial discharge characteristic parameter array.

[0021] Based on the monitoring point array, a monitoring point data processing module set is constructed. Each monitoring point corresponds to a data processing module, which is responsible for processing and analyzing the partial discharge signal collected at the point. The monitoring point data processing module set is used to process the signal data of different points in parallel. Through the distributed computing method, a large amount of data can be processed at the same time, the computing efficiency can be improved, and the partial discharge signal of each monitoring point can be independently analyzed and feature extracted.

[0022] Distributed signal analysis specifically includes: each module performs time domain analysis on the received partial discharge signal, such as calculating the amplitude, time interval, waveform and other characteristics of the signal; each module performs frequency domain analysis on the signal through fast Fourier transform technology, extracts information such as frequency components and energy spectrum; based on the results of time domain and frequency domain analysis, each module further extracts characteristic parameters of partial discharge, such as discharge amplitude, repetition frequency, phase distribution, pulse shape, etc. Through distributed signal analysis, the data processing modules at each monitoring point can independently extract the characteristic parameters of the partial discharge signal they are responsible for, and summarize these characteristic parameters to form a complete partial discharge characteristic parameter array.

[0023] A partial discharge feature tree is constructed, and an array of partial discharge cause feature parameters is obtained based on the association matching between the partial discharge feature parameter array and the partial discharge feature tree.

[0024] The feature tree is a hierarchical data structure used to store and represent the relationship between partial discharge features and their potential causes. The nodes of the partial discharge feature tree represent partial discharge features, and the edges represent the associations between features. The vertex of the feature tree is a node representing partial discharge, which branches out to primary and secondary feature nodes, until the leaf nodes represent specific partial discharge causes. Based on historical data, partial discharge feature parameters and related discharge cause information are collected, and the association between each feature parameter and their relationship with the discharge cause are analyzed. Based on the results of feature association analysis, the hierarchical structure of the feature tree is defined. The top-level node is the partial discharge feature, which is decomposed into more detailed features layer by layer, and finally reaches the specific discharge cause node to construct the partial discharge feature tree.

[0025] The local discharge characteristic parameter array is associated and matched with the local discharge characteristic tree, that is, each feature in the local discharge characteristic parameter array is matched with the corresponding node on the characteristic tree by comparison and calculation, so as to find the most likely cause of the discharge. Specifically, each characteristic parameter in the local discharge characteristic parameter array is compared with the nodes at each level of the characteristic tree to find the best matching path. Similarity metrics such as cosine similarity can be used for matching. The best path in the characteristic tree is selected according to the matching results, that is, starting from the root node, along the matching conditions of the nodes at each level, and finally reaching one or more leaf nodes, that is, the discharge cause nodes. According to the terminal node of the best path, the possible cause of the local discharge is inferred. Through the above-mentioned association matching process, the characteristic parameters of the cause of the local discharge can be determined, and these parameters are combined to form an array of characteristic parameters of the cause of the local discharge.

[0026] A partial discharge operation and maintenance instruction is triggered, and a partial discharge operation and maintenance strategy library is retrieved based on the partial discharge operation and maintenance instruction, wherein the partial discharge operation and maintenance strategy library is constructed based on historical partial discharge operation and maintenance records, and each record in the historical partial discharge operation and maintenance records includes historical partial discharge characteristic information, historical partial discharge cause characteristic information, and historical partial discharge defect operation and maintenance information.

[0027] After obtaining the partial discharge cause characteristic parameter array, the partial discharge operation and maintenance instruction is triggered. The instruction is a specific operation instruction to guide the operator or the automation system to perform maintenance, repair or preventive operation on the transformer. Based on the instruction, the corresponding strategy is retrieved from the partial discharge operation and maintenance strategy library. The operation and maintenance strategy library is a database built based on historical partial discharge operation and maintenance records. It stores a large amount of partial discharge characteristics, cause characteristics and defect operation and maintenance information. This information is organized based on historical experience and actual operation records.

[0028] Among them, historical partial discharge characteristic information refers to the partial discharge characteristic data recorded at a certain point in the past, including discharge amplitude, frequency, phase characteristics, etc. These data help identify past discharge patterns and intensities; historical partial discharge cause characteristic information describes the specific causes of past partial discharges, such as insulation aging, internal defects, external environmental changes, etc. This information helps to identify the potential causes of current partial discharges; historical partial discharge defect operation and maintenance information includes specific operation and maintenance measures taken for a certain partial discharge phenomenon, effect evaluation, subsequent monitoring results, etc. This information is the core part of the strategy library and provides a direct reference for current decision-making.

[0029] According to the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array, matching is traversed in the partial discharge operation and maintenance strategy library to obtain a target partial discharge operation and maintenance strategy set, and partial discharge early warning operation and maintenance processing is performed on the target transformer based on the target partial discharge operation and maintenance strategy set.

[0030] In the partial discharge operation and maintenance strategy library, all historical records similar to the current partial discharge characteristic parameters and cause characteristic parameters are retrieved, and the similarity between the current characteristic parameters and the historical records in the strategy library is calculated using a similarity measurement method, such as cosine similarity. The higher the similarity, the more the historical record conforms to the actual situation of the current partial discharge. According to the size of the similarity, the matching historical records are sorted, and the historical partial discharge defect operation and maintenance information corresponding to the record with the highest similarity is selected as the target partial discharge operation and maintenance strategy for the current point. All points are traversed to obtain the target partial discharge operation and maintenance strategy set.

[0031] Based on the target partial discharge operation and maintenance strategy set, partial discharge early warning operation and maintenance processing is performed on the target transformer, that is, based on the instructions in the strategy set, preventive and corrective operations are implemented. Preventive operations include adjusting the operating load, changing the cooling method, increasing the insulation monitoring frequency, etc., to reduce the frequency or intensity of partial discharge. Corrective operations include local insulation reinforcement, discharge point repair, replacement of damaged parts, etc., to prevent or mitigate the damage of partial discharge to the transformer.

[0032] Furthermore, if Figure 2 As shown, each record in the historical partial discharge operation and maintenance record also includes historical partial discharge defect information, and the method further includes:

[0033] Based on the historical partial discharge operation and maintenance records, extract the historical partial discharge feature information set, the historical partial discharge cause feature information set, and extract the associated historical partial discharge defect information set, wherein the historical partial discharge defect information set includes but is not limited to air gap defect information, bubble defect information, crack defect information, and impurity defect information; use the historical partial discharge feature information set and the historical partial discharge cause feature information set as input and the historical partial discharge defect information set as output to perform supervised learning and model testing to construct a partial discharge defect analyzer; based on the partial discharge defect analyzer, perform the partial discharge feature parameter array, The partial discharge cause characteristic parameter array is used for defect analysis to obtain a partial discharge defect information array; random abnormal interference factors are introduced, and based on the abnormal interference factors and in combination with the partial discharge defect information array, a risk assessment framework is used to perform potential negative effect analysis to generate a potential negative effect set, wherein the abnormal interference factors include environmental change factors, power grid operation status factors, and material property mutation factors; the potential negative effect set is sorted from high to low, and the operation and maintenance intervention priority of each partial discharge point in the partial discharge point set is determined according to the sorting result; the operation and maintenance intervention priority is added to the target partial discharge operation and maintenance strategy set.

[0034] The historical partial discharge operation and maintenance record is a historical record of partial discharge of transformers, which contains characteristic parameters, cause analysis, operation and maintenance measures of partial discharge events. Each record represents a past operation and maintenance instance. All relevant partial discharge event data are mined from the historical records to form a historical partial discharge characteristic information set and a historical partial discharge cause characteristic information set. The historical partial discharge characteristic information set contains all partial discharge characteristic parameters recorded in history. These parameters reflect the specific characteristics and manifestations of different partial discharge events. The historical partial discharge cause characteristic information set contains characteristic information of various causes of partial discharge in history. This information helps to understand the causes of partial discharge and provide reference for current discharge events. At the same time, all defect types related to partial discharge are identified from historical records, including air gap defects, bubble defects, crack defects, and impurity defects. The extracted defect information is sorted and summarized to construct a historical partial discharge defect information set.

[0035] The historical partial discharge feature information set and the historical partial discharge cause feature information set are used as input features, and the historical partial discharge defect information set is used as output labels. That is, given the discharge features and cause features, the model should predict the possible defect types.

[0036] Supervised learning is a machine learning method that builds a model capable of prediction by learning the mapping relationship between input data and output labels. Specifically, the prepared data set is divided into a training set and a test set. For example, the division ratio is set to 70% for training and 30% for testing. The model structure is built based on neural networks, etc. The model is trained using the training set, and the model parameters are adjusted to minimize the training error. After the model training is completed, the trained model is tested using the test set, and indicators such as the model's prediction accuracy are calculated. Based on the test and verification results, the model's hyperparameters are adjusted or features are reselected to further optimize the model performance. Through supervised learning and model verification, a partial discharge defect analyzer is finally constructed that can predict the type of internal defects of the transformer based on partial discharge characteristics and cause characteristics.

[0037] The local discharge feature parameter array and the local discharge cause feature parameter array are input into the defect analyzer. The analyzer model determines the possible defects corresponding to the current input features based on the mapping relationship learned during its training process, and outputs the prediction results as the local discharge defect information array.

[0038] Random abnormal interference factors that may affect the behavior of partial discharge are introduced. These factors will interfere with the normal operation of the transformer and increase the risk of partial discharge. These factors represent the abnormal conditions that the transformer may encounter in actual operation, including but not limited to environmental change factors, power grid operation status factors, and material performance mutation factors. Among them, environmental change factors such as changes in environmental conditions such as temperature, humidity, air pressure, and external electromagnetic interference may affect the insulation performance and electric field distribution of the transformer, thereby changing the behavior of partial discharge; power grid operation status factors such as voltage fluctuations, load changes, short circuit faults and other changes in power grid operation status may cause changes in the electrical stress of the transformer, resulting in an increase in the frequency and intensity of partial discharge; material performance mutation factors such as aging, deterioration or mechanical damage of insulating materials directly affect the insulation strength of the transformer and increase the possibility of partial discharge. The purpose of introducing these abnormal interference factors is to simulate the behavior of the transformer under different abnormal conditions and evaluate the risk level under these conditions.

[0039] The abnormal interference factors are cross-analyzed with the partial discharge defect information array to evaluate how the abnormal factors affect the manifestation and development of existing defects. Based on different abnormal interference factors, multiple scenarios of transformer partial discharge are simulated to evaluate the risk level in these scenarios. For example, the changes in partial discharge intensity that may be caused by existing crack defects in high temperature and high humidity environments are simulated. The risk assessment framework is applied, which is a tool for systematic risk assessment. Specifically, possible risk events are identified, such as specific abnormal factors leading to enhanced partial discharge. The probability of occurrence and potential impact of each risk event are analyzed. The probability of occurrence is such as the increase in partial discharge intensity caused by power grid fluctuations, and the potential impact is such as increased insulation damage, shortened transformer life, and increased electrical accidents. The potential impact of each abnormal situation is combined with its probability of occurrence to calculate a comprehensive risk score. The higher the risk score, the more serious the potential negative effect, which needs to be handled first. Through the above risk assessment process, a potential negative effect set is generated. This set contains the possible negative effects and their severity under various abnormal interference factors, providing data support for subsequent decision-making and emergency response.

[0040] According to the set of potential negative effects, the comprehensive risk score of each potential negative effect is extracted, and the risk scores are sorted from high to low. The higher the risk score, the greater the probability of occurrence of the effect and the more serious the impact, which requires priority treatment. The sorted potential negative effect list is mapped with each local discharge point, and the operation and maintenance intervention priority of each local discharge point is determined according to the sorting position of the negative effect. The priority can directly use the sorting position of the risk effect. The point with the highest priority means that it has the greatest risk and needs intervention as early as possible.

[0041] The operation and maintenance intervention priority of each local discharge point is added to the target operation and maintenance strategy set. The updated strategy set not only includes the processing method of each point, but also includes the priority processing order of each point. For example, for points with higher priority, maintenance plans need to be arranged in advance.

[0042] Furthermore, the interactive acquisition of a potential partial discharge point set of a target transformer and the preset monitoring point array according to the potential partial discharge point set include:

[0043] Interactively obtain the potential local discharge point set and determine a number of initial monitoring points; perform ultra-high frequency signal propagation simulation based on the initial monitoring points to determine N ultra-high frequency signal propagation blind areas and M ultra-high frequency signal propagation redundant areas, where N≥0 and M≥0; perform signal propagation coverage analysis based on the N ultra-high frequency signal propagation blind areas and the M ultra-high frequency signal propagation redundant areas to locate N newly added monitoring points and M reduced monitoring points; update the initial monitoring points based on the N newly added monitoring points and the M reduced monitoring points to determine the monitoring point array.

[0044] Collect the transformer design drawings, historical operation data, sensor monitoring records and other information to identify the potential partial discharge point set, which includes all the locations where partial discharge may occur inside the transformer. These locations can be determined based on equipment design features, historical operation data, expert experience, etc. Based on the potential partial discharge point set, determine several initial monitoring points.

[0045] Computer simulation technology is used to simulate the propagation process of UHF signals inside the transformer. The simulation is based on electromagnetic field theory and signal propagation model. During the simulation, the coverage of the signal at different monitoring points is analyzed to determine the propagation path and coverage range of the UHF signal inside the transformer. The simulation results show the intensity distribution of the signal at different monitoring points. According to the simulation results, N UHF signal propagation blind areas and M UHF signal propagation redundant areas are determined. The blind area is the area that the signal cannot cover, and the redundant area is the area where the signal overlaps excessively. The number of blind areas and redundant areas is N and M respectively, and N≥0, M≥0.

[0046] Based on the N blind areas of UHF signal propagation, the locations of N new monitoring points are determined. These new points should cover all blind areas as much as possible to ensure the comprehensiveness of monitoring. Based on the M redundant areas of UHF signal propagation, M reduced monitoring points are determined. These reduced points can minimize signal redundancy while still maintaining effective monitoring of local discharge areas.

[0047] N new monitoring points are added to the initial monitoring point set, and M reduced monitoring points are removed from the set, and finally an optimized monitoring point array is determined, which covers all important partial discharge areas, has uniform signal coverage, and has no obvious blind spots and redundant areas.

[0048] Furthermore, the method of arranging an ultra-high frequency sensor group based on the monitoring point array, performing dynamic discharge monitoring on the target transformer, and obtaining an ultra-high frequency signal timing array includes:

[0049] The initialized UHF sensor group is sequentially deployed to the monitoring point array, and a first monitoring point is extracted, wherein the first monitoring point is any point in the monitoring point array; a first UHF signal timing is acquired through a first UHF sensor at the first monitoring point; and so on, the monitoring point array is traversed to acquire the UHF signal timing array.

[0050] All UHF sensors are initialized, including functional testing and calibration, to ensure that each sensor is functioning properly before installation. After initialization, each sensor is installed in turn at each point in the monitoring point array, and any point in the monitoring point array is randomly selected as the first monitoring point for analysis.

[0051] Start the first UHF sensor at the first monitoring point and begin to collect the UHF signal inside the transformer in real time. Partial discharge events usually generate high-frequency electromagnetic waves. The sensor can capture these signals and record their intensity, frequency, phase and other information to generate the first UHF signal timing.

[0052] Traverse all points in the monitoring point array, repeat the signal acquisition process for each point until the data of all monitoring points are collected, and summarize the signal timing data of all monitoring points to form a complete ultra-high frequency signal timing array, which can reflect the partial discharge activity inside the entire transformer.

[0053] Furthermore, the method of performing distributed signal analysis on the partial discharge signal timing array through the monitoring point data processing module set to obtain a partial discharge characteristic parameter array includes:

[0054] Based on the partial discharge signal timing array, extract the second partial discharge signal timing of the second monitoring point, wherein the second monitoring point is any point in the monitoring point array, and the second monitoring point corresponds to a second monitoring point data processing module; through the second monitoring point data processing module, perform Fourier transform on the second partial discharge signal timing to generate a second signal frequency feature; based on the second signal frequency feature, perform time domain analysis on the second partial discharge signal timing to generate a second signal time domain analysis result; based on the second signal frequency feature, perform frequency domain analysis on the second partial discharge signal timing to generate a second signal frequency domain analysis result; determine a second partial discharge characteristic parameter according to the second signal time domain analysis result and the second signal frequency domain analysis result; and so on, traverse the partial discharge signal timing array to obtain a partial discharge characteristic parameter array.

[0055] A random point is selected from the monitoring point array as a second monitoring point and an analysis object, and signal timing data of the second monitoring point is extracted from the partial discharge signal timing array as a second partial discharge signal timing.

[0056] The second monitoring point data processing module is a module used to process the signal data of the second monitoring point. It has the functions of performing various processing and analysis on the signal. The data processing module is used to perform Fourier transform on the signal data. Fourier transform is a mathematical transformation that converts the time domain signal into a frequency domain signal to reveal the spectral characteristics of the signal. The frequency characteristics of the signal are extracted from the Fourier transform results, such as the main frequency, frequency distribution, power spectral density, etc. The frequency characteristics can reflect the main frequency components and energy distribution of the partial discharge signal.

[0057] Time domain analysis is a method for analyzing the changing characteristics of a signal in the time dimension. It usually focuses on the signal's amplitude, phase, pulse shape, zero crossing point and other characteristics. Based on the frequency characteristics, it identifies the time domain characteristics in the signal timing, for example, it identifies the signal's main pulse position, amplitude change trend, phase drift, etc., and calculates the signal's time domain characteristic parameters, such as the signal's peak amplitude, rise time, fall time, pulse width, repetition frequency, etc. These parameters can reflect the time-varying characteristics of the partial discharge signal, integrate all the extracted time domain features, and generate the second signal time domain analysis result.

[0058] Frequency domain analysis is used to reveal the main frequency components, harmonics, noise characteristics, etc. of the signal. Specifically, based on the frequency characteristics obtained by Fourier transform, the signal's spectrum is analyzed in more detail to identify the amplitude and phase of specific frequency components, focusing on the main frequency, subfrequency and changing trend of the local discharge signal, analyzing the harmonic components and noise characteristics in the signal, identifying the harmonic components of the local discharge signal, such as the second harmonic, third harmonic, etc., and their relative energy. Noise analysis helps to distinguish the discharge signal from the background noise, calculate the power spectral density of the signal, and determine the energy distribution of the signal at different frequencies. Power spectral density analysis can help identify the frequency components of the signal and the corresponding energy intensity, integrate all frequency domain analysis data, and generate the second signal frequency domain analysis result.

[0059] The time domain analysis result and the frequency domain analysis result are integrated to form the second partial discharge characteristic parameter.

[0060] Traverse all monitoring points in the partial discharge signal timing array, perform time domain and frequency domain analysis on the signal timing of each monitoring point, extract and calculate the corresponding characteristic parameters, and integrate the characteristic parameters of all monitoring points into a complete partial discharge characteristic parameter array. Each element in the array represents a characteristic parameter of a monitoring point.

[0061] Furthermore, the method for constructing the partial discharge feature tree includes:

[0062] Acquire local discharge characteristic factor information, wherein the local discharge characteristic factor information includes local discharge intensity, discharge frequency, discharge duration, spectrum characteristics, and time domain characteristics; retrieve a pre-constructed local discharge defect library, and perform associated topological identification on the local discharge defect library based on the local discharge characteristic factor information to obtain a local discharge characteristic defect library; define a local discharge defect top node, and classify local discharge causes of the local discharge characteristic defect library based on the local discharge defect top node to obtain a first-level local discharge cause node set; perform local discharge feature cascading analysis based on the first-level local discharge cause node set to construct the local discharge feature tree.

[0063] Obtain information on local discharge characteristic factors, which are used to describe the nature and behavior of local discharge. Specifically, they include local discharge intensity, which refers to the amount of electricity or current intensity of local discharge, which is usually related to the energy of the discharge and the degree of damage to the insulation; discharge frequency, which refers to the frequency of local discharge events, which can reflect the activity and nature of the discharge; discharge duration, which refers to the duration of a single local discharge event, which helps to identify the type of discharge, such as surface discharge, breakdown discharge, etc.; as well as spectral characteristics and time domain characteristics.

[0064] The partial discharge defect library is a database containing all known partial discharge defect types and their characteristic descriptions. This library is built based on a large amount of historical data and experimental results, covering various types of partial discharge defects and their manifestations. Load the defect library, obtain all known defect types and their characteristic descriptions, and perform association analysis with each record in the defect library based on the acquired characteristic factor information. By calculating similarity or other association measurement methods, identify the association between characteristic factors and specific defect types. Based on the association analysis results, construct a topological structure between characteristic factors and defect types. This structure can be in the form of a network diagram, representing the relationship between different characteristic factors and multiple defect types. Based on the topological identification results, integrate the relevant defect information and characteristic factor information into a new partial discharge characteristic defect library. This characteristic defect library reflects the association and relationship between partial discharge characteristics and defect types.

[0065] The top node of partial discharge defect is defined to represent the event of "partial discharge occurs". The top node is the starting point of the analysis and is used to aggregate and classify all the characteristics and causes related to partial discharge. All characteristic information related to the cause of partial discharge is extracted from the constructed partial discharge characteristic defect library. The partial discharge characteristic defect library is classified according to the characteristic information. The classification standard can be based on actual experience and statistical analysis results. For example, discharge events with similar characteristics are classified into the same cause type. According to the classification results, a set of first-level partial discharge cause nodes is generated. Each node corresponds to a specific partial discharge cause category, such as insulation aging, air gap defects, mechanical damage, environmental changes, etc.

[0066] Based on the set of primary local discharge cause nodes, the analysis is performed to identify the cascade relationship between each node. For example, insulation aging may cause local overheating, which in turn leads to local discharge. Based on the identified cascade relationship, the initial structure of the feature tree is designed. The top node is the local discharge occurrence node, the first-level node is the direct cause node, such as insulation aging, air gap defects, etc., the second-level node is a more specific cause or influencing factor, and so on, until all known features and causes are included in the feature tree. Finally, the local discharge feature tree is output as an important tool for analyzing the causes and characteristics of local discharge.

[0067] Furthermore, the method of performing cascade analysis of partial discharge characteristics based on the first-level partial discharge cause node set to construct the partial discharge characteristic tree includes:

[0068] The first-level local discharge cause node set is respectively subjected to local discharge feature correlation analysis to obtain a first-level node local discharge feature parameter set; based on the first-level node local discharge feature parameter set, local discharge cause subdivision association is performed to construct a second-level local discharge cause node set; the second-level local discharge cause node set is respectively subjected to local discharge feature correlation analysis until a K-level local discharge cause node set is obtained, wherein K≥2; based on the local discharge defect top node, the first-level local discharge cause node set, the second-level local discharge cause node set, until the K-level local discharge cause node set, cascade deep clustering is performed to obtain the local discharge feature tree.

[0069] Local discharge feature correlation analysis is used to identify the correlation between different feature parameters and local discharge cause nodes, helping to understand how these features are associated with specific causes. Specifically, statistical methods such as the Pearson correlation coefficient and the Spearman rank correlation coefficient are used to calculate the correlation between each feature parameter and each cause node, identify which feature parameters are most significant in characterizing a specific cause node, and extract the key feature parameters of each first-level cause node based on the correlation calculation results. The feature parameters include discharge intensity, frequency characteristics, time domain characteristics, spectrum energy, etc. The extracted feature parameters are integrated into a set of first-level node local discharge feature parameters. Each node has a corresponding set of feature parameters to describe the characteristics of the cause node.

[0070] Based on the set of local discharge characteristic parameters of the first-level nodes, each first-level cause node is deeply analyzed to identify which characteristic parameters can be used to further subdivide the cause nodes. The subdivision criteria can be based on the significance, characteristic pattern, statistical distribution, etc. of the characteristic parameters. For example, according to different discharge intensities and frequency distributions, insulation aging can be subdivided into thermal aging and chemical aging. According to the subdivision criteria, the first-level cause nodes are subdivided to generate a set of second-level local discharge cause nodes. Each newly generated second-level node represents a more specific subdivision cause of the first-level node.

[0071] Perform feature correlation analysis on the secondary cause node set, similar to the method in the previous steps, identify the feature parameters of each secondary node, further subdivide the cause, and generate a third-level partial discharge cause node set based on the feature correlation analysis results. Each third-level node represents a more specific subdivided cause of the secondary node. Continue to perform feature correlation analysis on the new third-level cause node set to generate a fourth-level cause node set. Repeat this process. When the preset subdivision level (K level) is reached or the node subdivision no longer generates new meaningful cause nodes, stop further subdivision and output a K-level partial discharge cause node set, K≥2.

[0072] Cascade deep clustering is a clustering analysis method that forms a multi-level tree structure by aggregating cause nodes at each level. Deep clustering not only considers the direct relationship between nodes, but also analyzes their hierarchical position and association pattern in the entire structure. Specifically, starting from the top node, cluster analysis is performed step by step. First, the first-level cause nodes are aggregated, then the second-level cause nodes, and so on. The result of each level of clustering is based on the cluster center or structure of the previous level. According to the clustering results, a partial discharge feature tree is constructed. The top node is the root node. Each level of cause nodes is connected and branched according to the cluster center and distance relationship to form a tree structure. The final partial discharge feature tree is output as an important tool for partial discharge analysis, fault diagnosis and decision support.

[0073] Furthermore, the method of traversing and matching in the partial discharge operation and maintenance strategy library according to the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array to obtain a target partial discharge operation and maintenance strategy set includes:

[0074] Extracting a third partial discharge characteristic parameter and a third partial discharge cause characteristic parameter of a third monitoring point, wherein the third monitoring point is any one of the monitoring point arrays; randomly extracting a first historical partial discharge operation and maintenance record from the partial discharge operation and maintenance strategy library, obtaining first historical partial discharge characteristic information, first historical partial discharge cause characteristic information, and first historical partial discharge defect operation and maintenance information; comparing the third partial discharge characteristic parameter with the first historical partial discharge characteristic information, and comparing the third partial discharge cause characteristic parameter with the first historical partial discharge cause characteristic information, and obtaining a first similarity according to the comparison result; and so on, traversing the partial discharge operation and maintenance strategy library to obtain multiple similarities; obtaining the historical partial discharge defect operation and maintenance information corresponding to the highest similarity among the multiple similarities, and generating a third target partial discharge operation and maintenance strategy for the third monitoring point; and so on, traversing the monitoring point array to obtain the target partial discharge operation and maintenance strategy set.

[0075] A random point is randomly selected from the monitoring point array as the third monitoring point and the analysis object, and the third partial discharge characteristic parameter and the third partial discharge cause characteristic parameter of the third monitoring point are respectively extracted from the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array.

[0076] The partial discharge operation and maintenance strategy library is accessed, and a record is randomly selected from the strategy library as the first historical partial discharge operation and maintenance record, and a comparative analysis is prepared. The first historical partial discharge operation and maintenance record includes first historical partial discharge characteristic information, first historical partial discharge cause characteristic information, and first historical partial discharge defect operation and maintenance information.

[0077] Compare the local discharge characteristic parameters of the third monitoring point with the local discharge characteristic information in the historical records, for example, compare the discharge intensity, frequency characteristics, time domain characteristics, etc.; compare the local discharge cause characteristic parameters of the third monitoring point with the cause characteristic information in the historical records, for example, compare the insulation aging degree, environmental conditions, etc.; select a suitable similarity calculation algorithm, such as Euclidean distance, cosine similarity, Pearson correlation coefficient, etc., calculate the feature similarity, and generate a first similarity after calculation, which indicates the similarity between the current discharge situation and the historical records.

[0078] Each historical record in the partial discharge operation and maintenance strategy library is traversed, and for each record, the aforementioned similarity calculation process is repeated to obtain multiple similarities.

[0079] Among all the calculated similarities, the highest similarity and its corresponding historical record are selected. The record with the highest similarity represents the historical case closest to the current partial discharge situation. The partial discharge defect operation and maintenance information is extracted from the historical record corresponding to the highest similarity. Based on the extracted historical operation and maintenance information, a third target operation and maintenance strategy suitable for the current third monitoring point is formulated.

[0080] Traverse all monitoring points, repeat the data extraction, similarity calculation and strategy generation process for each point, integrate the target operation and maintenance strategies of all monitoring points, obtain the target partial discharge operation and maintenance strategy set, and provide comprehensive decision support for partial discharge management of transformers.

[0081] In summary, the transformer partial discharge monitoring method based on ultra-high frequency technology provided in the embodiment of the present application has the following technical effects:

[0082] Determine the potential local discharge point set of the target transformer and preset the monitoring point array to ensure that all areas where local discharge may occur are monitored. By deploying an ultra-high frequency sensor group for dynamic discharge monitoring, an ultra-high frequency signal timing array can be obtained, thereby accurately capturing the dynamic changes of the local discharge signal; by deploying a multi-dimensional sensor group for multi-dimensional dynamic environmental monitoring, a multi-dimensional environmental signal timing array is obtained, and based on these environmental signals, the ultra-high frequency signal timing array is filtered and compensated, effectively filtering out environmental noise and interference factors, improving the accuracy and credibility of the local discharge signal, and ensuring the high precision of the monitoring results; by constructing a set of monitoring point data processing modules, a distributed signal analysis is performed on the local discharge signal timing array to obtain a local discharge characteristic parameter array. The distributed processing method improves the speed and efficiency of data analysis, making it possible to analyze a large amount of monitoring data in real time, thereby quickly identifying local discharge characteristics; based on local discharge characteristics The parameter array constructs a local discharge feature tree and performs association matching to obtain a local discharge cause feature parameter array. The construction and matching process of this feature tree can systematically represent the relationship between local discharge features and causes, which helps to deeply understand the causes and development processes of local discharge and support more accurate fault diagnosis. By triggering local discharge operation and maintenance instructions and calling the operation and maintenance strategy library built based on historical local discharge operation and maintenance records, it can traverse and match in the strategy library according to the current local discharge feature parameters and cause feature parameters, and automatically generate a target local discharge operation and maintenance strategy set. This process realizes the automation and intelligence of operation and maintenance decision-making, making fault handling more efficient and targeted. Based on the target local discharge operation and maintenance strategy set, the target transformer is subjected to local discharge early warning operation and maintenance. This early warning mechanism can detect and handle potential local discharge faults at an early stage, effectively prevent the occurrence of major faults, and improve the operating safety and reliability of the equipment.

[0083] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but will conform to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A transformer partial discharge monitoring method based on ultra-high frequency technology, characterized in that: The method comprises: Interactively obtain a potential partial discharge point set of a target transformer, preset a monitoring point array according to the potential partial discharge point set, deploy an ultra-high frequency sensor group based on the monitoring point array, perform dynamic discharge monitoring on the target transformer, and obtain an ultra-high frequency signal timing array; A multi-dimensional sensor group is arranged based on the monitoring point array, multi-dimensional dynamic environmental monitoring is performed on the target transformer, a multi-dimensional environmental signal timing array is obtained, and filtering compensation is performed on the ultra-high frequency signal timing array based on the multi-dimensional environmental signal timing array to obtain a partial discharge signal timing array; According to the monitoring point array, a monitoring point data processing module set is constructed, and a distributed signal analysis is performed on the partial discharge signal timing array through the monitoring point data processing module set to obtain a partial discharge characteristic parameter array; Constructing a partial discharge feature tree, and performing association matching between the partial discharge feature parameter array and the partial discharge feature tree to obtain a partial discharge cause feature parameter array; triggering a partial discharge operation and maintenance instruction, and calling a partial discharge operation and maintenance strategy library based on the partial discharge operation and maintenance instruction, wherein the partial discharge operation and maintenance strategy library is constructed based on historical partial discharge operation and maintenance records, and each record in the historical partial discharge operation and maintenance records includes historical partial discharge characteristic information, historical partial discharge cause characteristic information, and historical partial discharge defect operation and maintenance information; According to the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array, matching is traversed in the partial discharge operation and maintenance strategy library to obtain a target partial discharge operation and maintenance strategy set, and partial discharge early warning operation and maintenance processing is performed on the target transformer based on the target partial discharge operation and maintenance strategy set.

2. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1, characterized in that: Each record in the historical partial discharge operation and maintenance record also includes historical partial discharge defect information, and the method further includes: Based on the historical partial discharge operation and maintenance records, extract a historical partial discharge feature information set, a historical partial discharge cause feature information set, and extract an associated historical partial discharge defect information set, wherein the historical partial discharge defect information set includes air gap defect information, bubble defect information, crack defect information, and impurity defect information; Using the historical partial discharge feature information set and the historical partial discharge cause feature information set as input and the historical partial discharge defect information set as output, supervised learning and model verification are performed to construct a partial discharge defect analyzer; Based on the partial discharge defect analyzer, defect analysis is performed on the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array to obtain a partial discharge defect information array; Introducing random abnormal interference factors, based on the abnormal interference factors, combined with the partial discharge defect information array, applying the risk assessment framework to perform potential negative effect analysis, and generating a potential negative effect set, wherein the abnormal interference factors include environmental change factors, power grid operation status factors, and material performance mutation factors; Sorting the potential negative effect set from high to low, and determining the operation and maintenance intervention priority of each local discharge point in the local discharge point set according to the sorting result; The operation and maintenance intervention priority is added to the target partial discharge operation and maintenance strategy set.

3. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1 is characterized in that: The method of interactively obtaining a potential partial discharge point set of a target transformer and presetting a monitoring point array according to the potential partial discharge point set includes: Interactively obtaining the potential partial discharge point set and determining a number of initial monitoring points; Based on the several initial monitoring points, an ultra-high frequency signal propagation simulation is performed to determine N ultra-high frequency signal propagation blind areas and M ultra-high frequency signal propagation redundant areas, wherein N≥0, M≥0; According to the N UHF signal propagation blind areas and the M UHF signal propagation redundant areas, signal propagation coverage analysis is performed to locate N newly added monitoring points and M reduced monitoring points; According to the N newly added monitoring points and the M reduced monitoring points, the plurality of initial monitoring points are updated to determine the monitoring point array.

4. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1, characterized in that: The method of arranging an ultra-high frequency sensor group based on the monitoring point array, performing dynamic discharge monitoring on the target transformer, and obtaining an ultra-high frequency signal timing array includes: The initialized ultra-high frequency sensor group is sequentially arranged in the monitoring point array, and a first monitoring point is extracted, wherein the first monitoring point is any point in the monitoring point array; By using the first ultra-high frequency sensor at the first monitoring point, a first ultra-high frequency signal timing is acquired; By analogy, the monitoring point array is traversed to obtain the UHF signal timing array.

5. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1, characterized in that: The method of performing distributed signal analysis on the partial discharge signal timing array through the monitoring point data processing module set to obtain a partial discharge characteristic parameter array includes: Extracting a second partial discharge signal timing of a second monitoring point based on the partial discharge signal timing array, wherein the second monitoring point is any one of the monitoring point arrays, and the second monitoring point corresponds to a second monitoring point data processing module; Performing Fourier transform on the second partial discharge signal time series through the second monitoring point data processing module to generate a second signal frequency feature; Performing a time domain analysis on the timing of the second partial discharge signal based on the frequency characteristics of the second signal to generate a time domain analysis result of the second signal; Performing frequency domain analysis on the second partial discharge signal timing based on the second signal frequency characteristics to generate a second signal frequency domain analysis result; Determining a second partial discharge characteristic parameter according to a time domain analysis result of the second signal and a frequency domain analysis result of the second signal; By analogy, the partial discharge signal timing array is traversed to obtain a partial discharge characteristic parameter array.

6. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1, characterized in that: The method for constructing a partial discharge feature tree comprises: Acquiring partial discharge characteristic factor information, wherein the partial discharge characteristic factor information includes partial discharge intensity, discharge frequency, discharge duration, spectrum characteristics, and time domain characteristics; Retrieving a pre-built partial discharge defect library, and performing associated topological identification on the partial discharge defect library based on the partial discharge characteristic factor information to obtain a partial discharge characteristic defect library; defining a local discharge defect top node, and classifying local discharge causes of the local discharge feature defect library based on the local discharge defect top node to obtain a first-level local discharge cause node set; A partial discharge feature cascade analysis is performed based on the first-level partial discharge cause node set to construct the partial discharge feature tree.

7. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 6, characterized in that: The method of performing cascade analysis of local discharge characteristics based on the first-level local discharge cause node set to construct the local discharge characteristic tree comprises: Performing partial discharge feature correlation analysis on the first-level partial discharge cause node set to obtain a first-level node partial discharge feature parameter set; Based on the set of primary node local discharge characteristic parameters, local discharge causes are subdivided and associated to construct a secondary local discharge cause node set; Performing partial discharge feature correlation analysis on the secondary partial discharge cause node sets respectively until a K-level partial discharge cause node set is obtained, where K≥2; Based on the local discharge defect top node, the first-level local discharge cause node set, the second-level local discharge cause node set, and up to the K-level local discharge cause node set, cascade deep clustering is performed to obtain the local discharge feature tree.

8. The transformer partial discharge monitoring method based on ultra-high frequency technology according to claim 1, characterized in that: The method of traversing and matching in the partial discharge operation and maintenance strategy library according to the partial discharge characteristic parameter array and the partial discharge cause characteristic parameter array to obtain a target partial discharge operation and maintenance strategy set comprises: Extracting a third partial discharge characteristic parameter and a third partial discharge cause characteristic parameter of a third monitoring point, wherein the third monitoring point is any one point in the monitoring point array; Randomly extracting a first historical partial discharge operation and maintenance record from the partial discharge operation and maintenance strategy library to obtain first historical partial discharge characteristic information, first historical partial discharge cause characteristic information, and first historical partial discharge defect operation and maintenance information; comparing the third partial discharge characteristic parameter with the first historical partial discharge characteristic information, comparing the third partial discharge cause characteristic parameter with the first historical partial discharge cause characteristic information, and obtaining a first similarity according to the comparison result; By analogy, the partial discharge operation and maintenance strategy library is traversed to obtain multiple similarities; Acquire historical partial discharge defect operation and maintenance information corresponding to the highest similarity among the multiple similarities, and generate a third target partial discharge operation and maintenance strategy for the third monitoring point; By analogy, the monitoring point array is traversed to obtain the target partial discharge operation and maintenance strategy set.

Images