A Transformer Dynamic Diagnosis and Early Warning System Based on Sulfide Deposition Characteristics

Abstract

The present invention discloses a transformer dynamic diagnosis and early warning system based on the deposition characteristics of sulfides, comprising: a monitoring unit for real-time monitoring of the sulfide deposition state inside the transformer to obtain the sulfide deposition characteristic parameters during the operation of the transformer; a first diagnosis unit for determining whether the current sulfide deposition characteristic parameters affect the insulation system to obtain a first diagnosis result; a second diagnosis unit for obtaining a second diagnosis result through the propagation and attenuation characteristics of electrical signals, thermal signals and vibration signals at different inter-turn discharge positions, types and intensities; a fault diagnosis unit for determining whether the insulation defects and inter-turn discharge phenomena inside the transformer reach the early warning level according to the first diagnosis result and the second diagnosis result; and an early warning unit for generating corresponding fault early warning information according to the current early warning level to remind users or operation and maintenance personnel in real time to take preventive maintenance or emergency disposal measures.

Description

Technical Field

[0001] The present invention relates to the technical field of transformers, and in particular to a transformer dynamic diagnosis and early warning system based on the deposition characteristics of sulfides. Background Art

[0002] A transformer is a crucial device in the power system, and its long-term safe and reliable operation is of great significance to the stability of the entire power grid. However, the insulating paper and insulating oil inside the transformer will deteriorate during long-term operation. Among them, the deposition of cuprous sulfide (Cu 2 2S) is considered to be one of the important reasons for transformer failure. The formation of cuprous sulfide is usually related to the reaction of sulfides in the insulating paper. This reaction will accelerate the aging of the insulating paper and lead to breakdown, thus triggering partial discharge or even full discharge phenomena, which greatly threatens the operation safety of the transformer.

[0003] Most of the existing detection technologies cannot achieve online and real-time monitoring. The internal deterioration process of the transformer is slow and not easily detectable, and the occurrence and development process of cuprous sulfide deposition are more concealed. Traditional technologies cannot effectively capture these microscopic changes in the early stage; the current technologies focus more on the diagnosis after a fault, lacking continuous monitoring and trend analysis of the deterioration process. Especially for the insulation failure risk caused by sulfide deposition, no forward-looking early warning information is provided; most technologies only rely on a single type of physical signal (such as electrical signal or thermal signal) to judge the insulation state, and fail to comprehensively analyze the changes of multiple physical signals (such as ultrasonic waves, electromagnetic signals, light radiation, etc.), resulting in insufficient understanding of the deterioration process.

[0004] Therefore, there is an urgent need for a transformer dynamic diagnosis and early warning system based on the deposition characteristics of sulfides. Summary of the Invention

[0005] The present invention provides a transformer dynamic diagnosis and early warning system based on the deposition characteristics of sulfides to solve the above problems existing in the prior art.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A transformer dynamic diagnosis and early warning system based on the deposition characteristics of sulfides, comprising:

[0008] A monitoring unit for real-time monitoring of the sulfide deposition state inside the transformer and obtaining sulfide deposition characteristic parameters during the operation of the transformer;

[0009] A first diagnosis unit for judging whether the current sulfide deposition characteristic parameters affect the insulation system and obtaining a first diagnosis result;

[0010] A second diagnosis unit, which is used to identify the specific location and type of the discharge fault and obtain the second diagnosis result by means of the propagation and attenuation characteristics of electrical signals, thermal signals and vibration signals at different inter-turn discharge positions, types and intensities;

[0011] A fault diagnosis unit, which is used to judge whether the insulation defect and inter-turn discharge phenomenon inside the transformer reach the warning level according to the first diagnosis result and the second diagnosis result;

[0012] A warning unit, which is used to generate corresponding fault warning information according to the current warning level and remind the user or the operation and maintenance personnel to take preventive maintenance or emergency disposal measures in real time.

[0013] Among them, it also includes:

[0014] A display terminal, which is used to display the sulfide deposition characteristic parameters at each monitoring time point, the detailed information of each potential fault point, the propagation path and attenuation characteristics of the inter-turn discharge signal, the fault type and the warning level, as well as the preventive or maintenance suggestions given.

[0015] Among them, the monitoring unit includes:

[0016] A transformer monitoring module, which is used to set multiple monitoring time points at preset time intervals during the operation of the transformer, and monitor the deposition states of the internal insulating paper and sulfide of the transformer in real time to obtain the sulfide deposition signals and operation state signals at each monitoring time point;

[0017] A signal processing module, which is used to filter out the noise in the sulfide deposition signal and the operation state signal to obtain the sulfide deposition characteristic data and process data, and the process data is the load current of the transformer;

[0018] A division module, which is used to align the sulfide deposition characteristic data and the process data, divide the working conditions according to the process data, and classify the sulfide deposition characteristic data under each working condition to obtain multiple groups of sulfide deposition data;

[0019] A characteristic parameter extraction module, which is used to extract the deposition thickness and distribution characteristics of sulfide on the insulating paper corresponding to each monitoring time point based on the deposition characteristic samples in each group of sulfide deposition data, and obtain the sulfide deposition characteristic parameters during the operation of the transformer.

[0020] Among them, the first diagnosis unit includes:

[0021] A diagnosis model construction module, which is used to align the obtained copper sulfide deposition characteristic parameters with the aging and deterioration state data of the transformer insulating paper to obtain an insulation system characteristic data set, and construct a first fault diagnosis model based on the insulation system characteristic data set;

[0022] Anomaly detection module, which is used to perform anomaly detection on the sulfide deposition characteristic data under each detection condition based on a fault diagnosis model, identify potential insulation defect points, evaluate the severity, and obtain the first diagnosis result.

[0023] Among them, obtaining the second diagnosis result includes:

[0024] Obtain multi-physical signal data at different inter-turn discharge positions, types, and intensities, and identify initial diagnosis data from the multi-physical signal data. Among them, the initial diagnosis data includes electrical signal parameters, thermal signal parameters, and vibration signal parameters;

[0025] According to the electrical signal parameters, thermal signal parameters, and vibration signal parameters, perform hierarchical processing on the initial diagnosis data to obtain the data to be diagnosed, specifically including:

[0026] Obtain propagation path parameters from the electrical signal parameters, and obtain attenuation characteristic parameters from the thermal signal parameters and vibration signal parameters;

[0027] Obtain the attenuation characteristic parameters associated with each propagation path parameter as the data group to be diagnosed;

[0028] After sorting each data group to be diagnosed according to the propagation path parameters, store the data corresponding to each propagation path parameter and attenuation characteristic parameter into the corresponding data group to be diagnosed, and then obtain the data to be diagnosed;

[0029] Obtain the transformer fault diagnosis knowledge base, and obtain fault mode data from the transformer fault diagnosis knowledge base;

[0030] After performing noise elimination processing on the fault mode data, construct a data feature vector, and obtain fault correlation data from the transformer fault diagnosis knowledge base;

[0031] After clustering the data feature vector according to the fault correlation data, obtain the data group to be trained, and train the initial model with the data group to be trained to obtain the second fault diagnosis model;

[0032] Input the data to be diagnosed into the preset second fault diagnosis model one by one to obtain the specific position and fault type of the discharge fault;

[0033] Generate the second diagnosis result according to the specific position and fault type of the discharge fault and the initial diagnosis data.

[0034] Among them, judging whether the insulation defect and inter-turn discharge phenomenon inside the transformer reach the warning level includes:

[0035] According to the first diagnosis result and the second diagnosis result, classify and warn the severity of the insulation defect and inter-turn discharge phenomenon. The warning levels are divided into the following four categories:

[0036] Warning level 0, no obvious insulation defects or turn-to-turn discharge phenomena, and the transformer is in normal operation;

[0037] Warning level 1, slight insulation defects or turn-to-turn discharge phenomena are detected. It is recommended to monitor regularly and record the change trend;

[0038] Warning level 2, moderate insulation defects or turn-to-turn discharge phenomena exist. It is recommended to take preventive measures and conduct in-depth detection of relevant areas;

[0039] Warning level 3, the insulation system or turn-to-turn discharge phenomenon has reached a serious level, with a relatively high risk. It is recommended to stop the machine for maintenance immediately.

[0040] Among them, corresponding fault warning information is generated, including:

[0041] The fault warning model outputs the current warning level according to the input first diagnosis result and second diagnosis result. When the fault warning model outputs the current warning level, obtain the fault warning information corresponding to the warning level;

[0042] According to the current warning level, obtain the user reminder strategy corresponding to the warning level, and generate corresponding fault warning information;

[0043] If the warning level is 0, generate a prompt message to inform that the equipment is operating normally;

[0044] If the warning level is 1, generate a suggestion message to remind the user to monitor regularly and record the change trend;

[0045] If the warning level is 2, generate a preventive maintenance suggestion to remind the user or the operation and maintenance personnel to conduct in-depth detection of relevant areas to prevent the expansion of the fault;

[0046] If the warning level is 3, generate an emergency shutdown and maintenance suggestion to remind the user to stop the machine immediately and conduct emergency maintenance;

[0047] Display the fault warning information in real time on the display terminal;

[0048] Obtain the preset fault auxiliary decision-making model, input the current warning level and the generated fault warning information into the fault auxiliary decision-making model, and obtain at least one maintenance or disposal suggestion information output by the auxiliary decision-making model;

[0049] According to the maintenance or disposal suggestion information, provide the user or the operation and maintenance personnel with specific maintenance plans and strategy bases to help the user or the operation and maintenance personnel conduct fault prevention or emergency disposal.

[0050] Among them, the display terminal includes:

[0051] The first display module is used to display the sulfide deposition trend at each monitoring time point on the time axis in a graphical manner according to the obtained sulfide deposition characteristic parameters, and display the sulfide deposition changes in the monitoring area;

[0052] The second display module is used to mark in detail the geographical locations, fault degrees and fault types of each potential fault point in a preset distributed fault point map;

[0053] The third display module is used to display the propagation path and attenuation characteristics of the inter-turn discharge signal, including the initial position, propagation direction and intensity change of the signal, and visualize it using color gradient or line thickness;

[0054] The fourth display module is used to identify the warning levels of each fault type using different colors or symbols according to the obtained warning level information, and classify and display each fault type in a list form;

[0055] The fifth display module is used to display the corresponding prevention or maintenance suggestions according to the obtained fault information and warning level, and associate and display the suggestion information with the corresponding fault point or monitoring time point.

[0056] Among them, a first fault diagnosis model is constructed based on the insulation system characteristic data set, including:

[0057] For the obtained copper sulfide deposition characteristic parameters of the transformer, combined with the aging and deterioration state data of the transformer insulating paper under each working condition, an insulation system characteristic data set is constructed;

[0058] Based on the insulation system characteristic data set, analyze the correlation between copper sulfide deposition and insulation paper breakdown phenomenon, judge whether the copper sulfide deposition affects the transformer insulation system, further determine the severity of the insulation defect, and mark potential fault points;

[0059] For each group of working condition data in the insulation system characteristic data set, construct a first fault diagnosis model, and perform fault identification on the insulation system characteristic data under each working condition based on the first fault diagnosis model to obtain a first diagnosis result, including:

[0060] For each group of insulation system characteristic data, select the copper sulfide deposition characteristic parameters and insulation paper aging and deterioration state data from it to form a sub-sample set, and initialize the diagnostic tree based on the sub-sample set;

[0061] Randomly select multiple features from the sub-sample set, determine multiple splitting nodes based on each feature, and split the sub-sample set into multiple subsets based on each splitting node;

[0062] Calculate the variance of the insulation system characteristic data corresponding to each splitting node, and determine the variance reduction coefficient of each splitting node;

[0063] For each piece of feature data, calculate the distance between the feature data and other feature data, and select the nearest neighbor feature data for clustering to obtain a clustering coefficient;

[0064] Normalize the variance reduction coefficient and the clustering coefficient, and set the weight ratio to adjust the contribution ratio of the two to form the final splitting decision index;

[0065] Determine the target splitting node based on the maximum splitting decision index, and recursively split the target node until reaching the leaf node of the diagnostic tree;

[0066] For each piece of feature data, calculate the path parameter in the diagnostic tree, and mark the feature data whose path parameter exceeds the preset threshold as potential fault points;

[0067] Analyze the potential fault points, and through spectrum analysis or related diagnostic means, combined with the characteristic parameters of copper sulfide deposition and the data of the aging deterioration state of insulating paper, judge the fault type and location of the insulation system.

[0068] Among them, obtaining the specific location and fault type of the discharge fault includes:

[0069] Input the first data to be diagnosed into the second fault diagnosis model, perform similarity matching on the first data to be diagnosed in the second fault diagnosis model, and output the current diagnosis result according to the matching result;

[0070] Starting from the second data to be diagnosed, input the current diagnosis result into the second fault diagnosis model together to obtain the corresponding diagnosis result, and integrate the first diagnosis result and all diagnosis results to obtain the specific location and fault type of the discharge fault.

[0071] Compared with the prior art, the present invention has the following advantages:

[0072] A transformer dynamic diagnosis and early warning system based on the sulfide deposition characteristics, comprising: a monitoring unit for real-time monitoring of the sulfide deposition state inside the transformer to obtain the sulfide deposition characteristic parameters during the operation of the transformer; a first diagnosis unit for determining whether the current sulfide deposition characteristic parameters affect the insulation system to obtain a first diagnosis result; a second diagnosis unit for identifying the specific location and type of the discharge fault by the propagation and attenuation characteristics of the electrical signal, thermal signal and vibration signal at different inter-turn discharge positions, types and intensities to obtain a second diagnosis result; a fault diagnosis unit for judging whether the insulation defect and inter-turn discharge phenomenon inside the transformer reach the early warning level according to the first diagnosis result and the second diagnosis result; an early warning unit for generating corresponding fault early warning information according to the current early warning level to remind the user or the operation and maintenance personnel to take preventive maintenance or emergency disposal measures in real time. It greatly improves the accuracy of fault diagnosis and early warning and reduces the situation of misjudgment and missed report.

[0073] Other features and advantages of the present invention will be described in the following specification, and in part will be obvious from the specification, or will be understood by implementing the present invention.

[0074] The technical solution of the present invention will be further described in detail below through the drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification, and are used to explain the present invention together with the embodiments of the present invention, and do not constitute a limitation to the present invention. In the drawings:

[0076] Figure 1 is a structural diagram of a transformer dynamic diagnosis and early warning system based on the sulfide deposition characteristics in an embodiment of the present invention;

[0077] Figure 2 is a structural diagram of the monitoring unit in an embodiment of the present invention;

[0078] Figure 3 is a structural diagram of the first diagnosis unit in an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0079] The following describes the preferred embodiments of the present invention with reference to the drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention and are not used to limit the present invention.

[0080] The embodiment of the present invention provides a transformer dynamic diagnosis and early warning system based on the sulfide deposition characteristics, comprising:

[0081] A monitoring unit for real-time monitoring of the sulfide deposition state inside the transformer to obtain the sulfide deposition characteristic parameters during the operation of the transformer;

[0082] The first diagnosis unit is used to determine whether the current sulfide deposition characteristic parameters affect the insulation system and obtain the first diagnosis result;

[0083] The second diagnosis unit is used to identify the specific location and type of the discharge fault by the propagation and attenuation characteristics of electrical signals, thermal signals and vibration signals at different inter-turn discharge positions, types and intensities, and obtain the second diagnosis result;

[0084] The fault diagnosis unit is used to judge whether the insulation defects and inter-turn discharge phenomena inside the transformer reach the warning level according to the first diagnosis result and the second diagnosis result;

[0085] The warning unit is used to generate corresponding fault warning information according to the current warning level and remind the user or the operation and maintenance personnel to take preventive maintenance or emergency disposal measures in real time.

[0086] The working principle of the above technical solution is as follows: The monitoring unit monitors the sulfide deposition state inside the transformer in real time. Sulfide deposition is usually caused by chemical reactions of the insulating oil inside the transformer at high temperatures to generate sulfides, which are deposited on the conductor surface or insulating material, affecting the insulation performance of the transformer. Through the monitoring unit, the system can continuously collect the sulfide deposition characteristic parameters during the operation of the transformer. These parameters include deposition rate, deposit thickness, sulfide content, etc.

[0087] The first diagnosis unit is responsible for analyzing the sulfide deposition characteristic parameters provided by the monitoring unit and judging whether these depositions affect the insulation system of the transformer. It determines whether the deposits will cause a decline in insulation performance through set thresholds or trend analysis.

[0088] The second diagnosis unit focuses on identifying the inter-turn discharge fault inside the transformer. It determines the specific location, type and intensity of the fault by analyzing the electrical signals, thermal signals and vibration signals generated during inter-turn discharge. This unit collects data through sensors at different positions and uses the propagation and attenuation characteristics to distinguish the signal sources.

[0089] The fault diagnosis unit summarizes the results of the first diagnosis unit and the second diagnosis unit, and comprehensively judges whether the insulation defects and inter-turn discharge phenomena inside the transformer have reached the warning level. If sulfide deposition and abnormal inter-turn discharge are detected at the same time, the fault diagnosis unit will generate a higher fault severity determination.

[0090] Based on the analysis results of the fault diagnosis unit, the warning unit generates corresponding warning information. If the diagnosis results indicate that the insulation defect or turn-to-turn discharge has reached a dangerous level, the warning unit will automatically send an alarm to the user or the operation and maintenance personnel, suggesting preventive maintenance or emergency disposal measures. The warning levels are usually divided into three levels: low, medium, and high, and different levels correspond to different response measures.

[0091] The beneficial effects of the above technical solution are as follows: Through the real-time monitoring and multi-dimensional signal analysis (electrical, thermal, vibration signals) of the monitoring unit, the potential fault types and locations inside the transformer can be accurately identified, greatly improving the accuracy of fault diagnosis and warning, and reducing the situations of misjudgment and missed reports; Early identification of potential hazards such as sulfide deposition and turn-to-turn discharge allows maintenance personnel to take preventive maintenance measures before the problem worsens, thus avoiding excessive wear and unplanned downtime of the transformer and extending the service life of the equipment; The automated diagnosis and warning system reduces the need for manual inspections and the burden on operation and maintenance personnel. At the same time, through the timely reminder of the intelligent warning system, operation and maintenance personnel can respond to potential problems faster, shortening the fault downtime of the equipment.

[0092] In another embodiment, it further includes:

[0093] A display terminal, which is used to display the sulfide deposition characteristic parameters at each monitoring time point, the detailed information of each potential fault point, the propagation path and attenuation characteristics of the turn-to-turn discharge signal, the fault type and warning level, as well as the preventive or maintenance suggestions given.

[0094] The working principle of the above technical solution is as follows: The core function of the display terminal is to collect and display real-time data from various monitoring units, diagnostic units, fault diagnosis units, and warning units. The monitoring system of the transformer will monitor data such as sulfide deposition and inter-turn discharge in real time through sensors, and transmit this data to the display terminal through a communication network. The display terminal will display the sulfide deposition characteristic parameters at each monitoring time point in the form of charts or numbers. These parameters include the thickness of the deposit, the deposition rate, and the change trend over time. Users can intuitively view the data through the graphical interface of the terminal, facilitating real-time understanding of the operating status of the equipment. When the diagnostic unit detects potential fault points, the display terminal will display detailed information about these fault points, including location, signal characteristics, fault type, and development trend. This information can help the operation and maintenance personnel quickly understand the severity of the problem and its possible impacts. By analyzing the propagation path and attenuation characteristics of the discharge signal, the display terminal can display the signal propagation process in the form of a dynamic schematic diagram. For example, information such as how the signal propagates through different windings or insulation layers after leaving the discharge point, and how the signal intensity gradually attenuates with distance or time. This helps to locate the fault point and evaluate the scope of the fault impact. According to the analysis results of the fault diagnosis unit, the display terminal will display the identified fault types, such as inter-turn discharge, insulation aging, etc., and their corresponding warning levels (low, medium, high). The warning levels are classified according to the severity of the fault and the level of risk, helping users to make quick decisions. The display terminal not only displays fault information, but also provides specific preventive or maintenance suggestions based on the system analysis results. The suggestions may include measures such as equipment inspection, component replacement, or adjustment of operating conditions, to help users deal with potential problems in a timely manner. For example: When the display terminal detects moderate sulfide deposition, the system will prompt "It is recommended to replace the insulating oil or perform a cleaning treatment" to prevent the deposit from further deteriorating the insulation performance of the equipment.

[0095] The beneficial effects of the above technical solution are as follows: The display terminal helps the operation and maintenance personnel to grasp the operating status of the transformer in real time by converting complex data into intuitive charts, dynamic animations, and warning information. Whether it is sulfide deposition, inter-turn discharge, or fault location, the terminal can present them in a clear form, reducing the threshold for understanding complex technologies; The display terminal can display detailed information about potential faults and warning levels in real time, helping the operation and maintenance personnel to quickly identify problems and take measures. Through the efficient reminder of the warning system, the operation and maintenance personnel can respond before the problem develops to a serious stage, reducing the downtime of the transformer due to faults.

[0096] In another embodiment, the monitoring unit includes:

[0097] The transformer monitoring module is used to set multiple monitoring time points at preset time intervals during the operation of the transformer, real-time monitor the deposition status of the internal insulating paper and sulfide in the transformer, and obtain the sulfide deposition signals and operating status signals at each monitoring time point;

[0098] The signal processing module is used to filter out the noise in the sulfide deposition signals and operating status signals, obtain the sulfide deposition characteristic data and process data, and the process data is the load current of the transformer;

[0099] The partitioning module is used to align the sulfide deposition characteristic data and process data, divide the working conditions according to the process data, and classify the sulfide deposition characteristic data under each working condition to obtain multiple groups of sulfide deposition data;

[0100] The characteristic parameter extraction module is used to extract the deposition thickness and distribution characteristics of sulfide on the insulating paper corresponding to each monitoring time point based on the deposition characteristic samples in each group of sulfide deposition data, and obtain the sulfide deposition characteristic parameters during the operation of the transformer.

[0101] The working principle of the above technical solution is as follows: After the transformer starts to operate, the monitoring module sets multiple monitoring time points at preset time intervals (such as every 1 hour). The monitoring module can real-time monitor the deposition status of the internal insulating paper and sulfide in the transformer. The information obtained at these monitoring time points mainly includes two parts: one is the sulfide deposition signal, and the other is the operating status signal of the transformer.

[0102] Sulfide deposition signal: This is the original data collected from the internal sensors of the transformer regarding the deposition degree of sulfide on the insulating paper. Operating status signal: This is the data related to the operation of the transformer, such as load current, etc.

[0103] When the monitoring data is collected, this data contains some noise. The task of the signal processing module is to process this data and filter out the noise. In this way, the sulfide deposition characteristic data and process data can be obtained more accurately. If external electromagnetic interference or sensor self-error generates noise, the signal processing module will remove this noise through specific algorithms. Sulfide deposition characteristic data: After processing, the data reflecting the deposition amount of sulfide on the insulating paper is extracted. Process data: The change situation of the transformer load current, etc. These data help to understand the relationship between the working load of the transformer and sulfide deposition.

[0104] Among the obtained data, the sulfide deposition characteristic data and process data are aligned. The partitioning module classifies this data according to the operating conditions of the transformer. Working condition partitioning: According to the differences in the load current or other process parameters of the transformer, several different working conditions are divided. For example, low-load operation, high-load operation, over-load operation, etc.

[0105] Classification data: Under different working conditions, the deposition characteristics of sulfides will vary. This module classifies and stores the deposition data under each working condition. In this way, multiple sets of sulfide deposition data can be obtained, facilitating subsequent analysis.

[0106] Finally, the system extracts characteristic parameters based on each set of sulfide deposition data. By analyzing each set of deposition characteristic samples, the deposition thickness and distribution characteristics of sulfides on the insulating paper can be obtained. Deposition thickness and distribution characteristics: For example, at a certain moment, the sulfide deposition thickness in a certain area of the insulating paper is 0.2 mm, and the distribution is relatively uniform, or the deposition thickness in a certain area is large, which may indicate local aging problems. Deposition characteristic parameters: Based on this information, the system can generate a set of deposition characteristic parameters to reflect the deposition law of sulfides in the transformer under different operating states.

[0107] The beneficial effects of the above technical solution are as follows: Through this system, the deposition state of sulfides inside the transformer can be monitored in real time, avoiding potential problems such as the aging of insulating paper caused by excessive deposition, and reducing the risk of equipment failure; the signal processing module improves the accuracy of sulfide deposition characteristic data by filtering out noise. In this way, subsequent analysis is more reliable, which helps to detect the deterioration of the insulating paper inside the transformer in advance; by classifying the sulfide deposition data under different operating conditions of the transformer through the classification module, the influence of different operating conditions on the aging of insulating paper can be understood more clearly, which helps to optimize the operation strategy of the transformer; by extracting key parameters such as the deposition thickness and distribution characteristics of sulfides, the internal operating state of the transformer can be intuitively understood and a guiding basis can be provided for maintenance personnel. For example, when the deposition thickness is too large, the system can issue a warning to prompt maintenance.

[0108] In another embodiment, the first diagnosis unit includes:

[0109] A diagnosis model construction module, which is used to align the obtained copper sulfide deposition characteristic parameters with the aging and deterioration state data of the transformer insulating paper to obtain an insulating system characteristic data set, and construct a first fault diagnosis model based on the insulating system characteristic data set;

[0110] An anomaly detection module, which is used to perform anomaly detection on the sulfide deposition characteristic data under each detection working condition based on the fault diagnosis model, identify potential insulation defect points, and evaluate the severity to obtain a first diagnosis result.

[0111] The working principle of the above technical solution is as follows: First, the system will obtain the extracted copper sulfide deposition characteristic parameters, such as the deposition thickness and distribution characteristics of sulfides. At the same time, the system also needs to collect data on the aging and deterioration status of the transformer's insulating paper, which are historical data or obtained through other monitoring devices. Before building the model, the system needs to align the copper sulfide deposition characteristic parameters with the aging status data of the insulating paper. That is to say, the two types of data need to be synchronized so that they correspond in the same time dimension or working conditions. For example, at the same monitoring time point, there should be both aging data of the insulating paper and data on sulfide deposition. After alignment, the system combines these data to form an insulating system characteristic data set, which contains the complete characteristics of copper sulfide deposition data and the aging status of the insulating paper under different operating conditions of the transformer.

[0112] Construction of the first fault diagnosis model: Based on this data set, the system will build a first fault diagnosis model through a machine learning model. This model can identify the impact of copper sulfide deposition on the insulating system under different conditions and predict possible faults in the insulating system. If the system finds that the thickness of sulfide deposition exceeds a certain threshold and the degree of deterioration of the insulating paper also increases, the model may judge this situation as a potential insulating system fault and predict that the fault may occur at a certain time in the future.

[0113] After the diagnostic model is constructed, the system will use this model to perform real-time anomaly detection on the subsequently monitored copper sulfide deposition characteristic data. The system detects the copper sulfide deposition data under current working conditions based on the diagnostic model and identifies any situations that do not conform to the normal range. For example, during a certain monitoring, the system finds that the deposition thickness of copper sulfide suddenly increases, and the model predicts that this increase in deposition may lead to accelerated aging of the insulating paper. Then the system will mark this situation as abnormal. By analyzing the abnormal data, the system can identify possible defect points in the insulating system. For example, excessive concentration of sulfide deposition in a certain area may indicate more serious aging of the insulating paper in that area and even possible local damage. Not only identifying defects, the system will also evaluate the severity of the defects based on the diagnostic model. For example, excessive deposition may accelerate the time to insulation failure. The system can inform the operation and maintenance personnel of the urgency of the problem through prediction and provide corresponding maintenance suggestions. Finally, the system will generate a first diagnostic result, which details the current operating status of the transformer, the detected anomalies, and the recommended maintenance measures.

[0114] The beneficial effects of the above technical solutions are as follows: Through the construction of the diagnostic model, the system can accurately predict the impact of copper sulfide deposition on the aging of insulating paper, and identify potential fault points in advance, helping the transformer operation and maintenance team take measures in advance to avoid the occurrence of sudden failures; Through automated anomaly detection, the system can monitor the copper sulfide deposition situation during the operation of the transformer in real time, and promptly identify potential insulation defects, reducing the burden of manual inspection and greatly improving the detection efficiency; In addition to simply identifying faults, the system can also evaluate the severity of faults. This means that the operation and maintenance personnel can reasonably arrange the maintenance plan according to the system's suggestions, thus avoiding unnecessary shutdowns or repairs and optimizing the operation management of the transformer; Through the diagnostic model, the system can make maintenance decisions based on big data analysis according to the insulation system characteristic data. This method reduces the traditional experience-based maintenance mode, is more scientific and accurate, and thus reduces the operation and maintenance costs.

[0115] In another embodiment, obtaining the second diagnostic result includes:

[0116] Obtain multi-physical signal data at different inter-turn discharge positions, types, and intensities, and identify initial diagnostic data from the multi-physical signal data, where the initial diagnostic data includes electrical signal parameters, thermal signal parameters, and vibration signal parameters;

[0117] According to the electrical signal parameters, thermal signal parameters, and vibration signal parameters, perform hierarchical processing on the initial diagnostic data to obtain the data to be diagnosed, specifically including:

[0118] Obtain the propagation path parameters from the electrical signal parameters, and obtain the attenuation characteristic parameters from the thermal signal parameters and vibration signal parameters;

[0119] Obtain the attenuation characteristic parameters associated with each propagation path parameter as the data group to be diagnosed;

[0120] After sorting each data group to be diagnosed according to the propagation path parameters, store the data corresponding to each propagation path parameter and attenuation characteristic parameter into the corresponding data group to be diagnosed, and then obtain the data to be diagnosed;

[0121] Obtain the transformer fault diagnosis knowledge base, and obtain the fault mode data from the transformer fault diagnosis knowledge base;

[0122] After performing noise elimination processing on the fault mode data, construct a data feature vector, and obtain the fault association data from the transformer fault diagnosis knowledge base;

[0123] After clustering the data feature vector according to the fault association data, obtain the data group to be trained, and train the initial model with the data group to be trained to obtain the second fault diagnosis model;

[0124] Input the data to be diagnosed into a preset second fault diagnosis model one by one to obtain the specific location and fault type of the discharge fault;

[0125] Generate a second diagnosis result based on the specific location and fault type of the discharge fault and the initial diagnosis data.

[0126] The working principle of the above technical solution is as follows: First, tests are carried out at different inter-turn discharge positions (for example, inter-turn discharge between the primary winding and the secondary winding). In this process, various physical signal data will be collected, such as:

[0127] Electrical signal: Due to the change in the electric field, the discharge will generate instantaneous high-frequency voltage fluctuations and current pulses;

[0128] Thermal signal: Energy will be released during the discharge process, resulting in a local temperature increase;

[0129] Vibration signal: The shock wave generated by the discharge will cause slight vibrations in the physical structure of the transformer.

[0130] These signals are monitored in real time by different sensors, and the data is stored. This data constitutes the initial diagnosis data, mainly including:

[0131] Electrical signal parameters: such as voltage waveform, pulse frequency;

[0132] Thermal signal parameters: such as the rate of temperature change;

[0133] Vibration signal parameters: such as vibration frequency and intensity.

[0134] Next, perform hierarchical processing on the initial diagnosis data, that is, extract feature data at different levels according to different signal types:

[0135] Obtain propagation path parameters from electrical signals: This refers to the conduction path of the discharge current in the transformer winding or other components, such as the time of current propagation, path length, etc.;

[0136] Obtain attenuation characteristic parameters from thermal signals and vibration signals: These parameters describe the energy attenuation of thermal signals and vibration signals when transmitted in space, such as the ratio of temperature attenuation with distance, the attenuation rate of vibration amplitude, etc.

[0137] Through this hierarchical processing, the key information in the multi-physical signal data can be extracted to form "data to be diagnosed".

[0138] For each propagation path parameter, the system will search for the associated attenuation characteristic parameter and combine them into a set of data to be diagnosed. For example, if the propagation path of a certain discharge is short and the attenuation is small, then the system will mark its data separately, and then sort these data to finally generate a complete data set to be diagnosed.

[0139] Extract historical fault data from the fault diagnosis knowledge base of the transformer. The knowledge base contains rich fault modes and fault correlation data, such as the location, type of known discharge faults, and corresponding signal characteristics. By performing noise elimination processing on these fault modes, data feature vectors are generated.

[0140] The system classifies these feature vectors through a clustering algorithm to form a "data group to be trained". Next, these data groups to be trained are input into the initial fault diagnosis model for training, and finally a second fault diagnosis model is obtained, which can more accurately identify the specific location and type of discharge faults.

[0141] During the diagnosis process, the system will input the data to be diagnosed one by one into the trained second fault diagnosis model, and the model will identify the specific location and type of discharge faults based on the input data. For example:

[0142] The system identifies that the discharge occurs in the primary winding of the transformer;

[0143] The discharge type is partial discharge, and the intensity is high-frequency short-time discharge.

[0144] Combined with the initial diagnosis data, the system generates a second diagnosis result and feeds it back to the user.

[0145] The beneficial effects of the above technical solution are as follows: By integrating the parameters of various physical signals such as electricity, heat, and vibration, the characteristics of discharge faults can be captured more comprehensively and accurately, avoiding misjudgment caused by a single signal; The mechanism of hierarchical processing of electrical signals, thermal signals, and vibration signals enables the diagnostic system to extract key information more efficiently. In particular, the extraction and combination of propagation path parameters and attenuation characteristic parameters can enable the system to locate faults more quickly; The system can realize self-learning and optimization of the diagnostic model through the rich fault modes and historical data in the knowledge base. As more historical fault data is input, the diagnostic accuracy of the model will also continue to improve; By performing noise elimination processing on the fault mode data, the system can significantly improve the quality of diagnostic signals and classify faults through feature clustering, greatly reducing the probability of misjudgment; The data to be diagnosed can be quickly input into the model for real-time processing, and finally a second diagnosis result is generated. This real-time nature provides strong support for timely discovery and solution of potential faults in the transformer, avoiding the risk of fault expansion.

[0146] In another embodiment, determining whether the insulation defects and turn-to-turn discharge phenomena inside the transformer reach the warning level includes:

[0147] According to the first diagnosis result and the second diagnosis result, the severity of insulation defects and turn-to-turn discharge phenomena is classified and warned. The warning levels are divided into the following four categories:

[0148] Warning level 0, no obvious insulation defects or inter-turn discharge phenomena, and the transformer is in normal operation;

[0149] Warning level 1, slight insulation defects or inter-turn discharge phenomena are detected, and it is recommended to monitor regularly and record the change trend;

[0150] Warning level 2, moderate insulation defects or inter-turn discharge phenomena exist, and it is recommended to take preventive measures and conduct in-depth detection of relevant areas;

[0151] Warning level 3, the insulation system or inter-turn discharge phenomenon has reached a serious level, with a relatively high risk, and it is recommended to stop the machine for maintenance immediately.

[0152] The working principle of the above technical solution is as follows: First, the operating parameters of the transformer are collected through an on-line or off-line monitoring system, such as current, voltage, temperature, gas content in oil (by gas chromatography), and partial discharge (PD) data. These parameters can be used to evaluate the insulation status of the transformer.

[0153] Gas chromatography: By analyzing the gas components in the transformer oil, it is possible to detect the arc discharge and overheating phenomena inside the transformer. Gases such as hydrogen, acetylene, and methane are indicative gases of faults.

[0154] Partial discharge monitoring: By using sensors to detect the intensity and location of partial discharge phenomena, partial discharge is an important indicator of insulation defects, especially in the case of inter-turn discharge.

[0155] Preliminary diagnosis (the first diagnosis result): Based on the above data, a preliminary diagnosis is made to determine whether there are insulation defects or inter-turn discharge phenomena. If the monitoring data significantly exceeds the normal range, the preliminary diagnosis can give abnormal phenomena and prompt further analysis.

[0156] In-depth analysis (the second diagnosis result): Based on the results of the preliminary diagnosis, more refined analysis tools and algorithms (such as pattern recognition, trend analysis, etc.) are used to further confirm the severity of the problem. For example, the amplitude, frequency, and other characteristics of the partial discharge signal can be used to evaluate the severity of the insulation defect. At the same time, trend analysis is used to judge whether the problem is deteriorating.

[0157] Warning level 0: If the analysis results show that the data is completely within the normal range and there are no obvious abnormalities, it is judged that the transformer is in normal operation.

[0158] Warning level 1: If slight insulation defects or occasional inter-turn discharge phenomena are found in the analysis, the risk is relatively small, and the system recommends regular monitoring and recording of changes for continuous tracking.

[0159] Warning Level 2: When there are moderate insulation defects or obvious inter-turn discharge phenomena, it is recommended to take preventive measures such as reducing the load, improving the ventilation and cooling efficiency, and conducting a detailed inspection of the abnormal area to avoid further deterioration of the fault.

[0160] Warning Level 3: When the insulation system or inter-turn discharge phenomenon reaches a severe level, which means there is a high risk of short circuit or equipment failure, the equipment must be shut down immediately for maintenance to avoid greater losses or equipment damage.

[0161] Execution feedback: According to the warning classification results, the operation and maintenance personnel can take corresponding actions to ensure the safe and reliable operation of the transformer. The warning system will automatically record all operations and results, providing a basis for subsequent analysis.

[0162] The beneficial effects of the above technical solutions are as follows: Through timely detection and warning, measures can be taken before the insulation defects or inter-turn discharge problems deteriorate, preventing transformer failures or damages, and improving the safety and reliability of transformer operation; Through regular monitoring and preventive maintenance, the damages caused by insulation aging and inter-turn discharge can be reduced, thereby extending the service life of the transformer.

[0163] In another embodiment, corresponding fault warning information is generated, including:

[0164] The fault warning model outputs the current warning level according to the input first diagnosis result and second diagnosis result. When the fault warning model outputs the current warning level, obtain the fault warning information corresponding to the warning level;

[0165] According to the current warning level, obtain the user reminder strategy corresponding to the warning level, and generate corresponding fault warning information;

[0166] If the warning level is 0, generate a prompt message to inform that the equipment is operating normally;

[0167] If the warning level is 1, generate a suggestion message to remind the user to regularly monitor and record the change trend;

[0168] If the warning level is 2, generate a preventive maintenance suggestion to remind the user or the operation and maintenance personnel to conduct an in-depth inspection of the relevant area to prevent the expansion of the fault;

[0169] If the warning level is 3, generate an emergency shutdown and maintenance suggestion to remind the user to shut down immediately and conduct emergency maintenance;

[0170] Display the fault warning information in real time on the display terminal;

[0171] Obtain the preset fault auxiliary decision-making model, input the current warning level and the generated fault warning information into the fault auxiliary decision-making model, and obtain at least one maintenance or disposal suggestion information output by the auxiliary decision-making model;

[0172] Based on the maintenance or disposal suggestion information, provide users or operation and maintenance personnel with specific maintenance plans and strategic basis to help users or operation and maintenance personnel prevent faults or carry out emergency disposal.

[0173] The working principle of the above technical solution is as follows: the fault warning model first receives the operating data of the transformer, including the first diagnostic result and the second analysis result. The first diagnostic result is usually derived from conventional monitoring data, such as temperature, current, voltage, insulation status, etc. The second analysis result is obtained through in-depth analysis technology (such as trend analysis, pattern recognition, partial discharge monitoring, etc.), and contains more detailed fault information, such as insulation degradation trend, frequency and intensity of partial discharge, etc.

[0174] Based on the first diagnosis result and the second analysis result, the fault warning model determines the current warning level through a built-in algorithm. The warning levels are divided into four:

[0175] Warning level 0: No fault, the equipment operates normally.

[0176] Warning level 1: Minor insulation defect or mild inter-turn discharge, the equipment is in good condition.

[0177] Warning level 2: Moderate insulation defect or interturn discharge, the equipment is in poor condition and preventive maintenance is required.

[0178] Warning level 3: Severe insulation defects or severe inter-turn discharges, the equipment is at risk of major failure and must be shut down immediately.

[0179] Fault warning information generation: The model generates corresponding fault warning information according to different warning levels. The specific generation steps are as follows:

[0180] Warning level 0: Generates a warning message to inform that the device is operating normally and no action is required.

[0181] Warning Level 1: Generates advisory information to remind users or maintenance personnel to regularly monitor and record changing trends in equipment operation.

[0182] Warning level 2: Generates preventive maintenance recommendations to remind operation and maintenance personnel to conduct in-depth inspection and maintenance of the areas where problems are found to prevent the problems from further expanding.

[0183] Warning level 3: Generate emergency shutdown and maintenance recommendations, remind users to shut down immediately, and suggest arranging emergency maintenance to prevent equipment damage or shutdown accidents.

[0184] Real-time display of warning information: Fault warning information is displayed in real time on the display terminal, including the current warning level and corresponding operation suggestions, to help users and operation and maintenance personnel quickly understand the equipment status and take action.

[0185] Invocation of the auxiliary decision-making model: After generating the fault warning information, the system inputs the warning level and the generated fault warning information into a preset fault auxiliary decision-making model. The auxiliary decision-making model is a model based on data and empirical rules, and it will provide more specific repair or disposal suggestions according to the input data. For example, for the case of warning level 2, the auxiliary decision-making model will recommend detailed testing of the transformer winding in a specific area or replacement of the insulating material.

[0186] Generating repair suggestions: The auxiliary decision-making model outputs at least one repair or disposal suggestion. These suggestions include: specific repair operation steps (such as parts or materials to be replaced), the priority of the detection area (such as windings, insulating media, etc.), or emergency response measures (such as an emergency shutdown plan).

[0187] Finally, the system integrates the suggestion information of the auxiliary decision-making model into an actionable repair or prevention plan and displays it on the terminal for users or operation and maintenance personnel to refer to. These plans provide specific bases for the repair strategies for users and help them take corresponding measures to prevent the further deterioration of the fault.

[0188] The beneficial effects of the above technical solutions are as follows: By combining the first diagnosis result and the second analysis result, the system can more accurately evaluate the health status of the transformer and give early warnings of faults. This accurate early warning reduces the possibility of false alarms or missed alarms, enabling users to make more accurate repair decisions; Through hierarchical early warning and corresponding user reminder strategies, operation and maintenance personnel can take preventive measures in advance, effectively avoiding the expansion of potential faults and reducing the losses caused by equipment damage or fault shutdown; When a fault occurs, the system can quickly generate emergency handling suggestions and provide specific repair plans through the auxiliary decision-making model. This speeds up the problem-solving speed and shortens the downtime after equipment failure.

[0189] In another embodiment, the display terminal includes:

[0190] The first display module is used to display the sulfide deposition trend at each monitoring time point on the time axis in a graphical manner according to the obtained sulfide deposition characteristic parameters, and display the change of sulfide deposition in the monitoring area;

[0191] The second display module is used to mark in detail the geographical location, fault degree and fault type of each potential fault point in a preset distributed fault point map;

[0192] The third display module is used to display the propagation path and attenuation characteristics of the inter-turn discharge signal, including the initial position, propagation direction and intensity change of the signal, and visualize it by using color gradient or line thickness;

[0193] The fourth display module is used to, according to the obtained warning level information, use different colors or symbols to identify the warning levels of each fault type, and classify and display each fault type in the form of a list;

[0194] The fifth display module is used to, according to the obtained fault information and warning level, display corresponding prevention or maintenance suggestions, and associate and display the suggestion information with the corresponding fault point or monitoring time point.

[0195] The working principle of the above technical solution is as follows: The first display module (sulfide deposition trend display) obtains the sulfide deposition characteristic parameters at multiple monitoring time points, such as sulfide concentration, deposition rate, etc., correlates these data with time, and displays the sulfide deposition trend on the time axis in a graphical manner (such as line chart, bar chart, etc.). After the values at each monitoring time point are processed by an algorithm, a trend curve is generated to show the change of sulfide deposition. The second display module (distributed fault point location and display) first obtains potential fault point information, such as geographical location, fault severity, and fault type, through a sensor network or other monitoring devices. This information is combined with a preset distributed fault point map, and the specific location of the fault point is marked through a graphical interface, and colors, symbols, or other marks are used to indicate the fault degree (for example, red represents severe, yellow represents medium, and green represents minor). For example: In a power grid monitoring system, the sensor detects multiple faults in the cable. The second display module marks them on the geographical distribution map according to the specific coordinates of each fault point, and uses marks of different colors to represent the severity of the faults.

[0196] The third display module (visualization of the propagation path of inter-turn discharge signals) captures the propagation information of inter-turn discharge signals, including the initial position, propagation direction, and intensity change of the signals. Through data such as the propagation path and attenuation rate of the signals, this module uses color gradients (from bright to dark) or line thicknesses (from thick to thin) to visually display the change of signal intensity. For example: In the windings of a transformer, an inter-turn discharge phenomenon is detected. The third display module generates a line diagram with color gradients according to the discharge signals collected by the sensors, showing the process of the discharge signal starting from the discharge point and propagating along the windings and gradually attenuating. The line color changes from red to blue, and the line becomes thinner from thick, indicating the gradual weakening of the signal intensity.

[0197] The fourth display module (fault warning level display) identifies according to different warning levels (such as minor, warning, severe) using different colors or symbols based on the warning level information obtained from the fault monitoring system. At the same time, the system will display the types of all current faults in a list form and classify them according to the warning level. For example, the high-level warnings are ranked first in the list to remind the operator to handle them first. For example: In the power system, when the temperature of a transmission line is too high, the system issues a warning signal. The fourth display module uses a red triangle symbol to indicate "severe warning" and displays this fault at the top of the warning list to remind the relevant personnel to take immediate action. The fifth display module (prevention and maintenance suggestion display) generates corresponding prevention or maintenance suggestions by analyzing the obtained fault information and warning level. These suggestions will be displayed associatively according to different fault points or monitoring time points. For example, if the temperature is detected to be too high at a certain fault point, the system will provide corresponding cooling measure suggestions; if the sulfide deposition exceeds the warning value within a certain time point, the system will recommend relevant treatment measures. The suggestion content can be displayed in the form of graphics, text, etc. and associated with specific fault points or time points. When the bearing temperature of a certain generator is too high, the system generates a maintenance suggestion, indicating that the component needs to be lubricated immediately. The fifth display module displays this suggestion in a pop-up window on the monitoring interface and marks it at the bearing fault point to remind the maintenance personnel to pay attention to this problem.

[0198] The beneficial effects of the above technical solution are as follows: It helps to monitor the changes of sulfide deposition in real time, helps the relevant departments understand the emission trend of pollution sources, adjust environmental protection measures in time, and reduce the risk of environmental pollution; it can quickly locate potential fault points and evaluate the severity of faults, greatly improving the efficiency of fault troubleshooting, helping maintenance personnel to react quickly and reducing system downtime; through the intuitive display of the signal propagation path, technicians can more clearly understand the influence range and intensity change of the discharge phenomenon, which helps to accurately judge the cause of equipment faults and take targeted measures.

[0199] In another embodiment, a first fault diagnosis model is constructed based on the insulation system characteristic data set, including:

[0200] For the obtained copper sulfide deposition characteristic parameters of the transformer, combined with the aging and deterioration state data of the transformer insulating paper under each working condition, an insulation system characteristic data set is constructed;

[0201] Based on the insulation system characteristic data set, analyze the correlation between copper sulfide deposition and insulation paper breakdown phenomenon, judge whether the copper sulfide deposition affects the transformer insulation system, further determine the severity of insulation defects, and mark potential fault points;

[0202] For each set of operating condition data in the insulation system characteristic dataset, construct a first fault diagnosis model, and based on the first fault diagnosis model, identify faults in the insulation system characteristic data under each operating condition to obtain a first diagnosis result, including:

[0203] For each set of insulation system characteristic data, select copper sulfide deposition characteristic parameters and insulation paper aging and deterioration state data from it to form a sub-sample set, and initialize a diagnostic tree based on the sub-sample set;

[0204] Randomly select multiple features from the sub-sample set, determine multiple splitting nodes based on each feature, and split the sub-sample set into multiple subsets based on each splitting node;

[0205] Calculate the variance of the insulation system characteristic data corresponding to each splitting node, and determine the variance reduction coefficient of each splitting node;

[0206] For each characteristic data, calculate the distance between the characteristic data and other characteristic data, and select the nearest neighbor characteristic data for clustering to obtain a clustering coefficient;

[0207] Normalize the variance reduction coefficient and the clustering coefficient, and set a weight ratio to adjust the contribution ratio of the two to form a final splitting decision index;

[0208] Determine the target splitting node based on the maximum splitting decision index, and recursively split the target node until reaching the leaf node of the diagnostic tree;

[0209] For each characteristic data, calculate the path parameter in the diagnostic tree, and mark the characteristic data whose path parameter exceeds the preset threshold as potential fault points;

[0210] Analyze the potential fault points, and through spectrum analysis or related diagnostic means, combined with copper sulfide deposition characteristic parameters and insulation paper aging and deterioration state data, judge the fault type and location of the insulation system.

[0211] The working principle of the above technical solution is as follows: During the operation of the transformer, the deposition of copper sulfide will affect the insulation performance of the insulation paper and oil. Therefore, first, it is necessary to collect copper sulfide deposition characteristic parameters (such as copper sulfide deposition thickness, concentration, etc.) and transformer insulation paper aging and deterioration state data (such as dielectric loss, water content, fiber strength, etc.) under multiple operating conditions. These data are summarized to form a complete insulation system characteristic dataset for subsequent analysis. Suppose under a certain operating condition, the copper sulfide deposition concentration during the operation of the transformer is 10 ppm, and the dielectric loss of the insulation paper is 0.5%, and the water content is 2%. These parameters constitute a part of the characteristic data.

[0212] Based on the insulation system characteristic data set, through statistical analysis and correlation analysis, study whether the deposition of copper sulfide has a significant impact on the breakdown phenomenon of insulating paper, and further judge whether the deposition of copper sulfide will lead to a decline in insulation performance. By analyzing 100 groups of data, it is found that when the copper sulfide concentration exceeds 15 ppm, the dielectric loss of the insulating paper increases significantly and the breakdown voltage decreases significantly, indicating that there is a significant positive correlation between the deposition of copper sulfide and the deterioration of the insulating paper.

[0213] Based on the above data, a fault diagnosis model is constructed. First, select the copper sulfide deposition characteristic parameters and the insulating paper aging and deterioration state data as the sub-sample set, and initialize the diagnostic tree model. By randomly selecting multiple features in the sub-sample set (such as copper sulfide concentration, water content of insulating paper, etc.), multiple possible splitting nodes are determined. In a certain diagnosis, the copper sulfide concentration is selected as the feature, and the splitting node is set to 10 ppm, that is, when the copper sulfide concentration is higher or lower than 10 ppm, the health state of the insulation system is different. Through analysis, this node can effectively distinguish the normal and deteriorated insulation systems.

[0214] For the data corresponding to each splitting node, calculate its variance. The variance reduction coefficient can reflect the purity of the data set after splitting. At the same time, by calculating the distance between the feature data, feature clustering is performed, and the clustering coefficient is obtained. Through normalization processing and setting the weight ratio, the final splitting decision index is determined, and the splitting node corresponding to the maximum decision index is selected. In a certain calculation, the variance reduction coefficient of the splitting node of the copper sulfide concentration is 0.8, and the clustering coefficient is 0.9. After combining these two coefficients, the final splitting decision index is formed, and it is decided to split at the copper sulfide concentration of 15 ppm.

[0215] Through recursive splitting, a diagnostic tree model is finally constructed. The path of each feature data in the diagnostic tree will generate specific path parameters. For the feature data whose path parameters exceed the preset threshold, it is marked as a potential fault point. The result of a certain diagnosis shows that the copper sulfide deposition concentration exceeds 15 ppm, the dielectric loss exceeds 1%, and the path parameters exceed the normal value range, which is judged as a potential insulation system fault point.

[0216] Conduct a detailed analysis of the marked potential fault points. Diagnostic methods such as spectrum analysis and oil sample analysis can be used, combined with the copper sulfide deposition and insulating paper aging data, to further judge the fault type and location. For a certain potential fault point, through oil sample analysis, it is found that the copper sulfide concentration continues to rise. Combining with the aging data of the insulating paper, it is judged that the insulation system of this transformer is in a deteriorated state and is located at the A-phase winding.

[0217] The beneficial effects of the above technical solution are as follows: By constructing an insulation system feature dataset and a fault diagnosis model, the impact of copper sulfide deposition on the transformer insulation system can be analyzed more accurately. The diagnostic tree model can recursively split multiple nodes, making the fault identification process more refined, and ultimately accurately locating potential fault points; this model can monitor the health status of the insulation system in real time during the operation of the transformer, give early warnings about copper sulfide deposition and the aging of insulating paper, avoid sudden failures caused by insulation deterioration, and thus improve the overall operation reliability of the transformer; by discovering potential defects in the insulation system in advance, targeted maintenance can be carried out to avoid unnecessary shutdowns or replacement operations and reduce maintenance costs. At the same time, through the continuous optimization of the diagnostic model, the diagnostic strategy can be dynamically adjusted under different working conditions to improve the efficiency and accuracy of fault diagnosis.

[0218] In another embodiment, obtaining the specific location and type of the discharge fault includes:

[0219] Input the first data to be diagnosed into the second fault diagnosis model, perform similarity matching on the first data to be diagnosed in the second fault diagnosis model, and output the current diagnosis result according to the matching result;

[0220] Starting from the second data to be diagnosed, input the current diagnosis result into the second fault diagnosis model together to obtain the corresponding diagnosis result, and integrate the first diagnosis result and all diagnosis results to obtain the specific location and type of the discharge fault.

[0221] The working principle of the above technical solution is as follows: Suppose we have a set of data to be diagnosed, representing the sensor readings of a certain device, such as temperature, vibration frequency, etc. The first data to be diagnosed may be a temperature reading of 75°C and a vibration frequency of 50Hz.

[0222] Input the first data to be diagnosed into the second fault diagnosis model. This model is a machine learning-based algorithm, such as KNN (K-Nearest Neighbors), which is used to find the most similar record to the current data in the historical fault data. The model finds a similar record in the historical data, indicating that the device has experienced an overheating fault under similar conditions in the past.

[0223] According to the result of the similarity matching, the model outputs the current diagnosis result. For example: The diagnosis result is "overheating fault".

[0224] Starting from the second data to be diagnosed, input the current diagnosis result together with the new data to be diagnosed into the model. The second data is a vibration frequency of 55Hz, and it is input into the model in combination with the previous "overheating fault" result.

[0225] The model outputs updated diagnostic results based on new input data and previous diagnostic results. For example: The new diagnostic result is "overheating and abnormal vibration".

[0226] Integrate the first diagnostic result and all subsequent diagnostic results to determine the specific location and type of the fault. For example: The final diagnostic result is "There is overheating and abnormal vibration in the motor part of Device A".

[0227] The beneficial effects of the above technical solution are as follows: Through similarity matching, the model can use historical data to improve the accuracy of diagnosis and reduce misjudgment; With the input of new data, the model can update the diagnostic results in real time to help detect new fault characteristics in a timely manner; Integrate multiple diagnostic results to provide a more comprehensive fault analysis and help locate the specific location and type of the fault.

[0228] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and its equivalent technologies, the present invention is also intended to include these changes and modifications.

Claims

1. A transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics, characterized in that: include: A monitoring unit, used to monitor the sulfide deposition state inside the transformer in real time and obtain sulfide deposition characteristic parameters during transformer operation; A first diagnostic unit, used to determine whether the current sulfide deposition characteristic parameters have an impact on the insulation system and obtain a first diagnostic result; The second diagnosis unit is used to identify the specific location and fault type of the discharge fault through the propagation and attenuation characteristics of the electrical signal, thermal signal and vibration signal under different inter-turn discharge locations, types and intensities, and obtain a second diagnosis result; A fault diagnosis unit, used to determine whether the insulation defects and inter-turn discharge phenomena inside the transformer have reached a warning level according to the first diagnosis result and the second diagnosis result; The early warning unit is used to generate corresponding fault early warning information according to the current early warning level, and remind users or operation and maintenance personnel to take preventive maintenance or emergency disposal measures in real time; The first diagnostic unit includes: A diagnostic model building module, used to align the acquired copper sulfide deposition characteristic parameters with the aging and degradation state data of the transformer insulation paper, obtain an insulation system characteristic data set, and build a first fault diagnosis model based on the insulation system characteristic data set; An anomaly detection module is used to perform anomaly detection on the sulfide deposition characteristic data under each detection condition based on the first fault diagnosis model, identify potential insulation defect points, evaluate the severity, and obtain a first diagnosis result; The first fault diagnosis model is constructed based on the insulation system characteristic data set, including: Based on the obtained transformer copper sulfide deposition characteristic parameters, combined with the aging and degradation status data of the transformer insulation paper under each working condition, an insulation system characteristic data set is constructed; Based on the insulation system characteristic data set, the correlation between copper sulfide deposition and insulation paper breakdown is analyzed to determine whether copper sulfide deposition has an impact on the transformer insulation system, further determine the severity of insulation defects, and mark potential fault points; For each set of operating condition data in the insulation system characteristic data set, a first fault diagnosis model is constructed, and fault identification is performed on the insulation system characteristic data under each operating condition based on the first fault diagnosis model to obtain a first diagnosis result, including: For each set of insulation system characteristic data, copper sulfide deposition characteristic parameters and insulation paper aging degradation state data are selected to form a sub-sample set, and a diagnostic tree is initialized based on the sub-sample set; Randomly selecting multiple features from the sub-sample set, determining multiple splitting nodes based on each feature, and splitting the sub-sample set into multiple subsets based on each splitting node; Calculate the variance of the insulation system characteristic data corresponding to each split node, and determine the variance reduction coefficient of each split node; For each feature data, calculate the distance between the feature data and other feature data, and select the nearest neighbor feature data for clustering to obtain the clustering coefficient; The variance reduction coefficient and clustering coefficient are normalized, and the weight ratio is set to adjust the contribution ratio of the two to form the final split decision indicator; Determine the target split node based on the maximum split decision index, and recursively split the target node until the leaf node of the diagnosis tree is reached; For each feature data, the path parameter in the diagnosis tree is calculated, and the feature data whose path parameter exceeds the preset threshold is marked as a potential fault point; Analyze potential fault points, and determine the fault type and location of the insulation system by combining copper sulfide deposition characteristic parameters and insulation paper aging and degradation status data through spectrum analysis or related diagnostic methods.

2. A transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 1, characterized in that: Also includes: The display terminal is used to display the sulfide deposition characteristic parameters at each monitoring time point, detailed information of each potential fault point, the propagation path and attenuation characteristics of the inter-turn discharge signal, the fault type and warning level, as well as the prevention or maintenance suggestions given.

3. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 1 is characterized in that: The monitoring unit includes: The transformer monitoring module is used to set multiple monitoring time points according to preset time intervals during the operation of the transformer, monitor the deposition state of the insulation paper and sulfide inside the transformer in real time, and obtain the sulfide deposition signal and operation status signal at each monitoring time point; A signal processing module is used to filter out noise in the sulfide deposition signal and the operation status signal to obtain sulfide deposition characteristic data and process data, where the process data is the load current of the transformer; A division module is used to align the sulfide deposition characteristic data with the process data, divide the working conditions according to the process data, classify the sulfide deposition characteristic data under each working condition, and obtain multiple groups of sulfide deposition data; The characteristic parameter extraction module is used to extract the deposition thickness and distribution characteristics of sulfides on the insulating paper corresponding to each monitoring time point based on the deposition characteristic samples in each group of sulfide deposition data, and obtain the sulfide deposition characteristic parameters during the operation of the transformer.

4. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 1 is characterized in that: Obtain a second diagnosis, including: Acquire multi-physical signal data at different inter-turn discharge positions, types and intensities, and identify and obtain initial diagnostic data from the multi-physical signal data, wherein the initial diagnostic data includes electrical signal parameters, thermal signal parameters and vibration signal parameters; According to the electrical signal parameters, thermal signal parameters and vibration signal parameters, the initial diagnostic data is processed in layers to obtain the data to be diagnosed, including: Obtaining propagation path parameters from electrical signal parameters, and obtaining attenuation characteristic parameters from thermal signal parameters and vibration signal parameters; Acquire the attenuation characteristic parameter associated with each propagation path parameter as a data group to be diagnosed; After sorting each data group to be diagnosed according to the propagation path parameter, the data corresponding to each propagation path parameter and the attenuation characteristic parameter are stored in the corresponding data group to be diagnosed, thereby obtaining the data to be diagnosed; Acquire a transformer fault diagnosis knowledge base, and acquire fault mode data from the transformer fault diagnosis knowledge base; After noise elimination processing is performed on the fault mode data, a data feature vector is constructed, and fault-related data is obtained from the transformer fault diagnosis knowledge base; After clustering the data feature vectors according to the fault-related data, a data group to be trained is obtained, and after training the initial model with the data group to be trained, a second fault diagnosis model is obtained; Input the data to be diagnosed into the preset second fault diagnosis model one by one to obtain the specific location and fault type of the discharge fault; A second diagnosis result is generated according to the specific location and fault type of the discharge fault and the initial diagnosis data.

5. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 1 is characterized in that: Determine whether the insulation defects and inter-turn discharge phenomena inside the transformer have reached the warning level, including: According to the first and second diagnostic results, the severity of insulation defects and inter-turn discharge phenomena is graded and warned. The warning levels are divided into the following four categories: The warning level is 0, there is no obvious insulation defect or inter-turn discharge phenomenon, and the transformer is in normal operation; The warning level is 1, which means that a slight insulation defect or inter-turn discharge phenomenon is detected. It is recommended to monitor regularly and record the change trend. Warning level 2: moderate insulation defects or inter-turn discharges exist. It is recommended to take preventive measures and conduct in-depth inspections of relevant areas. The warning level is 3. The insulation system or inter-turn discharge has reached a serious level and there is a high risk. It is recommended to shut down the machine for maintenance immediately.

6. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 1 is characterized in that: Generate corresponding fault warning information, including: The fault warning model outputs a current warning level according to the input first diagnosis result and the second diagnosis result, and when the fault warning model outputs the current warning level, acquires fault warning information corresponding to the warning level; According to the current warning level, obtain the user reminder strategy corresponding to the warning level and generate corresponding fault warning information; If the warning level is 0, a prompt message is generated to inform that the device is operating normally; If the warning level is 1, a suggestion message is generated to remind the user to regularly monitor and record the change trend; If the warning level is 2, a preventive maintenance suggestion is generated to remind the user or operation and maintenance personnel to conduct in-depth inspections of the relevant areas to prevent the fault from expanding; If the warning level is 3, an emergency shutdown and maintenance suggestion is generated to remind the user to shut down immediately and perform emergency maintenance; Display fault warning information in real time on the display terminal; Obtaining a preset fault decision-making support model, inputting the current warning level and the generated fault warning information into the fault decision-making support model, and obtaining at least one maintenance or disposal suggestion information output by the decision-making support model; Based on the maintenance or disposal suggestion information, provide users or operation and maintenance personnel with specific maintenance plans and strategic basis to help users or operation and maintenance personnel prevent faults or carry out emergency disposal.

7. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 2 is characterized in that: The display terminal includes: The first display module is used to display the sulfide deposition trend at each monitoring time point on the time axis in a graphical manner according to the acquired sulfide deposition characteristic parameters, and to display the sulfide deposition changes in the monitoring area; The second display module is used to mark in detail the geographical location, fault degree and fault type of each potential fault point in a preset distributed fault point map; The third display module is used to display the propagation path and attenuation characteristics of the inter-turn discharge signal, including the initial position, propagation direction and intensity change of the signal, and is visualized by color gradient or line thickness; The fourth display module is used to identify the warning level of each fault type using different colors or symbols according to the acquired warning level information, and to display each fault type by category in the form of a list; The fifth display module is used to display corresponding prevention or maintenance suggestions according to the acquired fault information and warning level, and associate the suggestion information with the corresponding fault point or monitoring time point for display.

8. The transformer dynamic diagnosis and early warning system based on sulfide deposition characteristics according to claim 4 is characterized in that: Get the specific location and fault type of the discharge fault, including: Inputting the first data to be diagnosed into the second fault diagnosis model, performing similarity matching on the first data to be diagnosed in the second fault diagnosis model, and outputting the current diagnosis result according to the matching result; Starting from the second data to be diagnosed, the current diagnosis result is input into the second fault diagnosis model to obtain the corresponding diagnosis result, and the first diagnosis result and all the diagnosis results are integrated to obtain the specific location and fault type of the discharge fault.

Images