Non-contact large transformer aging detection method, system, device and medium

Abstract

The application provides a non-contact large transformer aging detection method, system, device and medium, relates to the technical field of transformer detection, and the method collects three-dimensional images, surface temperature distribution and environment data of the transformer, and utilizes GNSS synchronization time stamps to ensure data time sequence consistency. Based on the data, an accurate three-dimensional temperature field grid model is constructed, grid encryption processing is carried out for key areas, and an environment compensation coefficient is calculated to correct the influence of environmental factors. A dynamic threshold is learned through a prediction model, temperature anomaly features are analyzed and extracted from the three-dimensional temperature field grid model according to the dynamic threshold, and a machine learning model is used to evaluate the aging risk level, so that nondestructive, accurate monitoring and evaluation of the aging state of the large transformer are realized.

Description

Technical Field

[0001] This application relates to the field of transformer testing technology, and in particular to a non-contact method, system, device and medium for aging testing of large transformers. Background Technology

[0002] Large outdoor transformers are core hub equipment in power systems, and their operating status directly affects the stability of the regional power grid and the security of power supply. Outdoor large transformers have complex structures and experience severe load fluctuations, making them highly susceptible to external environmental influences, which can exacerbate problems and potentially lead to breakdowns or even explosions. These environmental factors include, but are not limited to, changes in temperature, humidity, and wind speed. For example, in high-temperature and high-humidity environments, the insulation materials inside the transformer may age more rapidly. Existing transformer aging detection technologies mainly include contact testing and manual testing. However, contact testing equipment is complex to install and has high maintenance costs, while manual testing is time-consuming and labor-intensive. These problems limit the effectiveness and reliability of traditional methods in the aging detection of large transformers. Summary of the Invention

[0003] This application provides a non-contact method, system, apparatus, and medium for aging detection of large transformers to solve one or more technical problems existing in the prior art, and at least provide a beneficial option or create conditions.

[0004] On the one hand, this application provides a non-contact aging detection method for large transformers, including the following steps:

[0005] The system acquires three-dimensional images, surface temperature distribution data, and environmental data of large transformers in a non-contact manner, and uses GNSS synchronization timestamps for time alignment.

[0006] Based on the three-dimensional image and the surface temperature distribution data, a three-dimensional temperature field mesh model of the large transformer is constructed, and the mesh of the winding region, core region and heat dissipation region is refined.

[0007] Based on the aforementioned environmental data and historical environmental data, the environmental compensation coefficient is calculated.

[0008] Based on the environmental compensation coefficient, the surface temperature distribution data, and the preset time sliding window, a dynamic threshold is calculated using a prediction model.

[0009] Based on the dynamic threshold, temperature anomaly features are extracted from the three-dimensional temperature field mesh model;

[0010] The abnormal temperature features are input into a machine learning model, which outputs the aging risk level of the large transformer.

[0011] Furthermore, the step of constructing a three-dimensional temperature field mesh model of the large transformer based on the three-dimensional image and the surface temperature distribution data, and refining the mesh of the winding region, core region, and heat dissipation region therein, includes the following steps:

[0012] Based on the three-dimensional image, an initial mesh structure model of the large transformer is generated using a three-dimensional reconstruction algorithm;

[0013] Using an image recognition algorithm, the winding region, the core region, and the heat dissipation region are identified and marked in the initial mesh structure model;

[0014] Each data point in the surface temperature distribution data is matched and mapped to the initial mesh structure model to construct a three-dimensional temperature field mesh model of the large transformer, and local mesh refinement processing is performed on the winding region, core region and heat dissipation region.

[0015] Furthermore, the environmental data includes real-time temperature, real-time humidity, and real-time wind speed, and the historical environmental data includes historical average temperature, historical average humidity, and historical average wind speed.

[0016] The environmental compensation coefficient satisfies the following formula:

[0017] ;

[0018] in, This represents the environmental compensation coefficient; This indicates the real-time temperature. This indicates the real-time humidity. This indicates the real-time wind speed; This represents the historical average temperature. This represents the historical average humidity. This represents the historical average wind speed; This represents the temperature weighting coefficient. This represents the humidity weighting coefficient. This represents the wind speed weighting coefficient; .

[0019] Furthermore, the dynamic threshold includes a temperature gradient amplitude threshold and a temperature standard deviation threshold;

[0020] The step of extracting temperature anomaly features from the three-dimensional temperature field mesh model based on the dynamic threshold includes the following steps:

[0021] The temperature gradient magnitude of each grid cell and the temperature standard deviation of adjacent grid cells in the three-dimensional temperature field grid model are calculated in real time.

[0022] Grid cells whose temperature gradient amplitude exceeds the temperature gradient amplitude threshold are marked as first-level abnormal regions;

[0023] Grid cells whose temperature standard deviation exceeds the temperature standard deviation threshold are marked as secondary anomaly regions;

[0024] The temperature gradient change rate of the primary anomaly region and the temperature fluctuation range of the secondary anomaly region are extracted as the temperature anomaly features.

[0025] Furthermore, the machine learning model includes a support vector machine model and a random forest model; the machine learning model adapts to different transformers through transfer learning.

[0026] Furthermore, the length of the time sliding window is 6 hours, and the sliding step is 30 minutes.

[0027] Furthermore, the prediction model includes a temporal convolutional network model; the temporal convolutional network model contains three dilated convolutional layers; wherein the dilation coefficient of the first dilated convolutional layer is 2, the dilation coefficient of the second dilated convolutional layer is 4, and the dilation coefficient of the third dilated convolutional layer is 8.

[0028] On the other hand, this application provides a non-contact aging detection system for large transformers, including a data acquisition module, a three-dimensional temperature field construction module, a temperature anomaly feature extraction module, and an aging risk prediction module.

[0029] The data acquisition module is used to collect three-dimensional images, surface temperature distribution data and environmental data of large transformers in a non-contact manner, and to perform time alignment using GNSS synchronization timestamps;

[0030] The three-dimensional temperature field construction module is used to construct a three-dimensional temperature field mesh model of the large transformer based on the three-dimensional image and the surface temperature distribution data, and to perform mesh refinement processing on the winding area, core area and heat dissipation area therein;

[0031] The temperature anomaly feature extraction module is used to calculate an environmental compensation coefficient based on the environmental data and historical environmental data; calculate a dynamic threshold based on the environmental compensation coefficient, the surface temperature distribution data, and a preset time sliding window through a prediction model; and extract temperature anomaly features from the three-dimensional temperature field mesh model based on the dynamic threshold.

[0032] The aging risk prediction module is used to input the abnormal temperature features into a machine learning model and output the aging risk level of the large transformer.

[0033] On the other hand, this application provides a non-contact large transformer aging detection device, which is mobile and deployed on unmanned inspection equipment, including a GNSS unit, a thermal imager, a 3D depth camera, an environmental sensor group, an edge computing unit and a communication unit;

[0034] The thermal imager is used to collect surface temperature distribution data of large transformers;

[0035] The 3D depth camera is used to acquire three-dimensional images of the large transformer;

[0036] The environmental sensor group is used to collect environmental data of the large transformer;

[0037] The GNSS unit is used to provide a synchronization timestamp to time-align the surface temperature distribution data, the three-dimensional image, and the environmental data.

[0038] The edge computing unit includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned non-contact large transformer aging detection method.

[0039] The communication unit is used to establish a communication connection with the remote monitoring platform.

[0040] On the other hand, this application provides a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to implement the aforementioned non-contact large transformer aging detection method.

[0041] The beneficial effects of this application are as follows: This application proposes a non-contact aging detection method for large transformers. This method acquires three-dimensional images of the transformer, surface temperature distribution, and environmental data, and uses GNSS synchronization timestamps to ensure data temporal consistency. Based on this data, a precise three-dimensional temperature field mesh model is constructed, and mesh refinement is applied to key areas. Simultaneously, an environmental compensation coefficient is calculated to correct for the influence of environmental factors. A dynamic threshold is learned through a predictive model, and temperature anomaly features are extracted from the three-dimensional temperature field mesh model based on the dynamic threshold. Then, a machine learning model is used to assess the aging risk level, thereby achieving non-destructive and accurate monitoring and assessment of the aging state of large transformers. This application also provides corresponding systems, devices, and media. The beneficial effects of the systems, devices, and media are similar to those of the method and will not be elaborated further here.

[0042] Other features and advantages of this application will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the application. The objectives and other advantages of this application may be realized and obtained by means of the structures particularly pointed out in the description, claims and drawings. Attached Figure Description

[0043] The accompanying drawings are provided to further understand the technical solutions of the present invention and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the technical solutions of the present invention, and do not constitute a limitation on the technical solutions of the present invention.

[0044] Figure 1 This is a flowchart of the non-contact aging detection method for large transformers provided in this application;

[0045] Figure 2 This is a schematic diagram illustrating the implementation process of the three-dimensional temperature field mesh model provided in this application;

[0046] Figure 3 This is a schematic diagram illustrating the process by which the aging risk level is obtained through testing, as provided in this application.

[0047] Figure 4 This is a structural diagram of the non-contact aging detection system for large transformers provided in this application;

[0048] Figure 5 This is a structural diagram of the non-contact large transformer aging detection device provided in this application. Detailed Implementation

[0049] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0050] The present application will be further described below with reference to the accompanying drawings and specific embodiments. The described embodiments should not be considered as limitations on the present application, and all other embodiments obtained by those skilled in the art without inventive effort are within the scope of protection of the present application.

[0051] In the following description, references are made to “some embodiments,” which describe a subset of all possible embodiments. However, it is understood that “some embodiments” may be the same subset or different subsets of all possible embodiments and may be combined with each other without conflict.

[0052] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0053] Large transformers are crucial components in power systems, and their operating status directly impacts the safety and stability of the power grid. Therefore, effective monitoring and assessment of the aging status of large transformers is of great significance. Existing aging detection methods mainly fall into two categories: contact and non-contact. Contact detection typically involves physically connecting sensors to critical parts of the transformer to obtain data. This method often requires shutdown operations, increasing power outage time and costs, and direct contact may cause potential damage to the equipment. Furthermore, due to limitations in sensor installation, contact detection struggles to achieve comprehensive monitoring of the entire equipment and cannot provide real-time, continuous data streams.

[0054] Moreover, large outdoor transformers are indeed significantly affected by numerous environmental factors due to the unique nature of their operating environment. Specifically: Outdoor transformers are directly exposed to the natural environment and experience huge temperature variations from cold to hot. Extreme temperature conditions not only affect the working efficiency of internal transformer components but also accelerate the aging process of insulation materials, increasing the risk of localized overheating, which may lead to equipment failure or shorten service life. High humidity environments can cause condensation inside the transformer, which may cause electrical short circuits or corrosion problems, especially for components that are not adequately sealed. Prolonged exposure to humid conditions can also weaken insulation performance, increase the likelihood of a drop in breakdown voltage, and threaten the safe and stable operation of the power grid.

[0055] Furthermore, while appropriate wind speeds aid in heat dissipation, strong winds or unstable airflow patterns can alter the effectiveness of the cooling system, resulting in some areas not receiving adequate cooling. Additionally, in severe weather conditions such as sandstorms, external particles entering the transformer can cause mechanical wear or other forms of damage.

[0056] On the other hand, while existing non-contact detection technologies can operate without interrupting service, most of these methods lack comprehensive analytical tools. For example, some technologies rely solely on a single type of sensor (such as an infrared thermal imager) to collect surface temperature information, failing to adequately integrate three-dimensional structural images and other environmental parameters for comprehensive analysis. This one-sided data acquisition approach may lead to inaccurate diagnostic results, making it impossible to accurately identify specific aging areas or types within the transformer. Furthermore, traditional methods perform poorly when dealing with complex environmental factors, such as failing to consider the impact of changes in ambient temperature and humidity on measurement results, which further affects the accuracy of aging condition assessment.

[0057] Secondly, in the evaluation method based on fixed thresholds, an alarm or shutdown operation is triggered when the temperature of a certain part of the transformer exceeds the preset safety limit. However, this method lacks the ability to predict the quantitative grading of aging and relies on a single indicator (such as temperature exceeding the limit) as the shutdown standard. It cannot effectively reflect the overall health status of the transformer, thus limiting its application scope and effectiveness.

[0058] Based on the above-mentioned situation, this application proposes a non-contact method, system, device, and medium for aging detection of large transformers. This application integrates multiple non-contact sensing technologies, including a 3D depth camera, a thermal imager, and an environmental sensor array, to simultaneously acquire the transformer's three-dimensional structure, surface temperature distribution, and operating environmental conditions. By constructing an accurate three-dimensional temperature field mesh model and utilizing machine learning algorithms, an accurate assessment of the transformer's aging degree is achieved, thus overcoming the shortcomings of existing technologies. This method improves the accuracy and efficiency of aging detection, reduces interference with equipment, and supports the stable operation of the power system.

[0059] First, the non-contact aging detection method for large transformers provided in this application will be described in detail below with reference to the accompanying drawings.

[0060] The non-contact aging detection method for large transformers proposed in this application can be applied to terminals, servers, or software running on either terminal or server. Terminals can be tablets, laptops, desktop computers, etc., but are not limited to these. Servers can be independent physical servers, server clusters or distributed systems composed of multiple physical servers, or cloud servers providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks, and big data and artificial intelligence platforms.

[0061] Reference Figure 1 The implementation process of the non-contact aging detection method for large transformers provided in this application includes, but is not limited to, the following steps.

[0062] Step S110: Collect three-dimensional images, surface temperature distribution data and environmental data of the large transformer in a non-contact manner, and perform time alignment using GNSS synchronization timestamps.

[0063] It should be noted that GNSS refers to the Global Navigation Satellite System. A GNSS synchronization timestamp refers to the generation of a unified time stamp for devices or systems by acquiring high-precision time signals from a global navigation satellite system (such as GPS, BeiDou, Galileo, etc.), enabling time synchronization across regions and devices. GNSS satellites carry highly stable atomic clocks and continuously broadcast navigation messages containing precise time information. Receiving terminals calculate the satellite signals to obtain the current Coordinated Universal Time (UTC) or system reference time, and then generate a timestamp. This timestamp has global consistency and nanosecond-level accuracy, and is widely used in scenarios requiring high-precision time synchronization, such as communication network synchronization, precision measurement, IoT device collaboration, and power system phase alignment, ensuring consistent time bases across all nodes and improving the accuracy of data recording and event sequencing.

[0064] In step S110, three-dimensional images, surface temperature distribution data, and environmental data of the large transformer are acquired using non-contact methods, and time-series alignment is performed using GNSS synchronization timestamps. This step provides fundamental data support for subsequent analysis, ensuring data temporal consistency and avoiding data analysis errors caused by time asynchrony. The non-contact acquisition method not only reduces operational interference with the equipment but also enables the acquisition of comprehensive data information without affecting the normal operation of the transformer.

[0065] Step S120: Based on the three-dimensional image and surface temperature distribution data, construct a three-dimensional temperature field mesh model of the large transformer, and refine the mesh of the winding area, core area and heat dissipation area.

[0066] It should be noted that the winding area is one of the core components of a large transformer, mainly composed of coils made of conductive materials such as copper or aluminum, responsible for transferring electrical energy between the primary and secondary sides. The operating condition of this area directly affects the electrical performance of the transformer; problems such as aging, overheating, or partial discharge can lead to decreased efficiency or even failure. Therefore, monitoring its temperature changes and insulation condition is particularly important.

[0067] The core region is typically composed of stacked silicon steel sheets, providing a low magnetic reluctance path to enhance the magnetic field and reduce energy loss. As the energy conversion center of the transformer, the efficient operation of the core is crucial for maintaining the stability and efficiency of the entire system. Any anomalies caused by overheating, magnetic saturation, or manufacturing defects will result in additional energy loss and may damage the windings and other components; therefore, this region is a key area for monitoring.

[0068] The heat dissipation area includes devices such as coolers, fans, and oil pumps, used to dissipate the heat generated by the transformer during operation and ensure that all components remain within a safe operating temperature range. Good heat dissipation design and functional effectiveness are crucial for extending transformer life and improving operating efficiency. If the heat dissipation system does not operate effectively, the internal temperature of the equipment will rise, accelerating the aging of insulation materials and increasing the risk of failure. Therefore, monitoring the status of the heat dissipation area helps to promptly identify and resolve potential temperature control problems.

[0069] In step S120, based on the acquired 3D images and surface temperature distribution data, a 3D temperature field mesh model of the large transformer is constructed, and the mesh is refined for the winding region, core region, and heat dissipation region. This step helps to more accurately simulate and analyze the internal temperature distribution of the transformer. Especially for critical parts, refining the mesh can improve the model's resolution, thereby more accurately capturing temperature changes in these areas and providing detailed thermodynamic basis for aging condition assessment.

[0070] Step S130: Calculate the environmental compensation coefficient based on environmental data and historical environmental data.

[0071] In step S130, an environmental compensation coefficient is calculated based on current and historical environmental data. Considering the influence of environmental factors such as temperature, humidity, and wind speed on the transformer surface temperature, this step corrects for these influencing factors by calculating the environmental compensation coefficient, ensuring the authenticity and accuracy of the temperature measurement results. The purpose of this is to eliminate the deviations caused by changes in the external environment to transformer temperature monitoring, making aging detection more reliable.

[0072] Step S140: Based on the environmental compensation coefficient, surface temperature distribution data, and preset time sliding window, the dynamic threshold is calculated using a prediction model.

[0073] In step S140, a dynamic threshold is calculated using an environmental compensation coefficient, surface temperature distribution data, and a preset time sliding window through a predictive model. This process aims to adjust the anomaly detection criteria according to real-time environmental conditions, making the detection criteria more flexible and adaptable. Setting a dynamic threshold can better identify abnormal temperature changes that may occur under specific environmental conditions, improving the sensitivity and accuracy of early fault warnings.

[0074] Optionally, the length of the time sliding window can be set to 6 hours, and the sliding step size can be set to 30 minutes.

[0075] Specifically, the 6-hour sliding window allows the system to capture the trend of transformer surface temperature changes over time and with the environment, while the 30-minute sliding step ensures highly sensitive monitoring of these trend changes. This setup not only identifies short-term temperature fluctuations but also effectively monitors long-term temperature change trends, thereby providing a more comprehensive assessment of the transformer's health status.

[0076] Step S150: Extract temperature anomaly features from the three-dimensional temperature field mesh model based on the dynamic threshold.

[0077] In step S150, temperature anomaly features are extracted from the three-dimensional temperature field mesh model based on dynamic thresholds. This step is crucial for achieving detailed analysis of the transformer's aging state. By calculating the temperature gradient amplitude and temperature standard deviation, potential aging hotspots or abnormal areas can be effectively identified. This anomaly detection method based on multi-dimensional features helps to accurately locate potential problems inside the transformer, laying the foundation for subsequent risk assessment.

[0078] Step S160: Input the temperature anomaly features into the machine learning model and output the aging risk level of the large transformer.

[0079] In step S160, the extracted temperature anomaly features are input into a machine learning model to output the aging risk level of the large transformer. This step applies a machine learning model, combined with a large amount of existing training data, to automatically assess the current aging condition of the transformer. This method not only improves diagnostic efficiency but also allows for adaptive adjustments based on different transformer characteristics, providing more personalized aging risk assessment results.

[0080] In some embodiments of this application, reference is made to Figure 2 In step S120, the process of constructing a three-dimensional temperature field mesh model of a large transformer based on the three-dimensional image and surface temperature distribution data, and performing mesh refinement processing on the winding region, core region and heat dissipation region therein, includes but is not limited to the following steps.

[0081] Step S210: Based on the three-dimensional image, generate the initial mesh structure model of the large transformer using a three-dimensional reconstruction algorithm.

[0082] In step S210, using the non-contact acquisition of 3D image data of the large transformer, a preliminary geometric mesh structure model is constructed using a 3D reconstruction algorithm. This model provides a spatial framework for subsequent temperature field mapping and analysis, ensuring a one-to-one correspondence between temperature distribution information and the actual physical structure. Establishing the initial mesh structure model is fundamental to constructing the 3D temperature field, helping to transform abstract data into a visualized physical model and providing support for refined analysis.

[0083] Step S220: Using an image recognition algorithm, identify and mark the winding region, core region, and heat dissipation region in the initial mesh structure model.

[0084] In step S220, based on the existing initial mesh model, key functional regions in the model are automatically identified and divided using image recognition techniques, such as convolutional neural networks (CNNs), edge detection algorithms, or feature matching algorithms. These regions include the core components of the transformer, namely the winding region, the core region, and the heat dissipation region. Accurately identifying and marking these regions provides a basis for subsequent local refinement and targeted analysis, improving the model's engineering practicality and diagnostic accuracy.

[0085] Step S230: Match and map each data point in the surface temperature distribution data to the initial mesh structure model to construct a three-dimensional temperature field mesh model of the large transformer, and perform local mesh refinement processing on the winding region, core region and heat dissipation region.

[0086] In step S230, the collected infrared thermal imaging or other forms of surface temperature distribution data are precisely superimposed onto the three-dimensional mesh structure model using coordinate mapping technology, thereby forming a three-dimensional temperature field mesh model containing temperature information. Based on this, local mesh refinement processing is performed on the identified key areas (winding area, core area, and heat dissipation area), that is, the mesh cells are further refined on the original mesh to improve the spatial resolution of the model. This refinement strategy can significantly improve the ability to capture temperature changes in key areas, helping to detect minute abnormal temperature rises, and thus more accurately reflect the aging state and potential fault risks of the transformer.

[0087] In some embodiments of this application, the 3D reconstruction algorithm employs a depth map fusion algorithm, such as the KinectFusion algorithm. This algorithm continuously captures depth images of the scene and fuses these images into a gradually refined 3D model in real time, thereby constructing an accurate geometric model of a large transformer. Depth map fusion algorithms such as KinectFusion are particularly suitable for non-contact 3D reconstruction tasks because they can not only handle complex object surfaces but also effectively reduce noise and voids, improving the quality and resolution of the reconstructed model. Using such algorithms for 3D modeling of transformers provides a high-precision spatial framework for subsequent steps such as temperature field analysis and aging assessment, ensuring the accuracy and reliability of data mapping. Furthermore, due to its efficient data processing capabilities, it can guarantee good reconstruction results even in dynamic environments, increasing the flexibility and applicability of the detection process.

[0088] In some embodiments of this application, environmental data includes real-time temperature, real-time humidity, and real-time wind speed, and historical environmental data includes historical average temperature, historical average humidity, and historical average wind speed. In step S130, the environmental compensation coefficient satisfies the following formula (1):

[0089] (1);

[0090] In formula (1), Indicates the environmental compensation coefficient; Indicates real-time temperature. Indicates real-time humidity. Indicates real-time wind speed; Indicates the historical average temperature. Indicates the historical average humidity. Indicates the historical average wind speed; This represents the temperature weighting coefficient. This represents the humidity weighting coefficient. This represents the wind speed weighting coefficient; .

[0091] Formula (1) calculates the environmental compensation coefficient by dynamically comparing real-time environmental data (temperature, humidity, wind speed) with historical averages and combining them with preset weighting coefficients, thereby accurately quantifying the impact of environmental factors on equipment operation. Specifically, firstly, dynamic calibration eliminates measurement deviations caused by environmental fluctuations, such as temperature measurement distortion in high-temperature and high-humidity environments; secondly, multi-factor collaborative compensation is implemented, including temperature weighting to dominate heat transfer, humidity weighting to correct insulation material performance degradation, and wind speed weighting to quantify changes in heat dissipation efficiency; furthermore, normalization ensures that the sum of weighting coefficients is 1, maintaining the consistency of the physical meaning of the compensation coefficients and allowing adjustment of weight combinations according to different regional climates; and thirdly, benchmark adaptation uses historical environmental averages as a reference system to identify the current environmental deviation relative to normal conditions. This compensation mechanism enables the aging detection system to eliminate environmental interference and focus on changes in the equipment's physical state, such as reducing the sensitivity of temperature-related parameters in high-temperature weather to avoid misjudgments, and rapidly responding to changes in heat dissipation conditions through wind speed weighting under strong wind conditions, thereby improving the reliability of temperature anomaly identification. Therefore, this dynamic compensation strategy is one of the core technical supports for achieving accurate aging detection.

[0092] In some embodiments of this application, a dynamic threshold is calculated using a prediction model based on an environmental compensation coefficient, surface temperature distribution data, and a preset time sliding window; wherein the prediction model includes a lightweight temporal model, such as a Temporal Convolutional Network (TCN).

[0093] TCN plays a crucial role in calculating dynamic thresholds based on environmental compensation coefficients, surface temperature distribution data, and a preset time sliding window. As a lightweight time-series model, TCN effectively captures long-term dependencies in time-series data. Through its unique dilated convolutional layers, it can significantly reduce the number of model parameters and computational complexity without sacrificing performance. This enables TCN to accurately identify trends and abnormal fluctuations in temperature changes in tasks requiring analysis of long-term data, such as aging detection of large transformers. This provides a reliable basis for calculating dynamic thresholds, enhancing the accuracy and timeliness of transformer condition assessment. Therefore, TCN not only improves the efficiency of the prediction model but also enhances the system's adaptability to environmental changes, making it an important tool for achieving accurate aging detection.

[0094] In some embodiments of this application, the TCN includes three dilated convolutional layers; wherein the dilation coefficient of the first dilated convolutional layer is 2, the dilation coefficient of the second dilated convolutional layer is 4, and the dilation coefficient of the third dilated convolutional layer is 8.

[0095] The first dilated convolutional layer allows the network to examine the time series of the input data at regular intervals (i.e., with a dilation factor of 2), meaning it can cover a relatively long time range without increasing the filter size. This design helps capture more recent but not instantaneous temperature change trends, and is particularly useful for identifying short-term fluctuations that may affect the health of transformers.

[0096] As the number of layers increases, the dilation coefficient of the second dilated convolutional layer increases to 4, which further expands the network's receptive field, enabling it to consider data variations over longer periods. This layer can identify trends or patterns that span hours or even longer, which is crucial for understanding how environmental factors affect the long-term operating condition of transformers.

[0097] The third dilated convolutional layer has a dilation coefficient of 8, which greatly expands the network's receptive field, allowing it to capture information on a more macroscopic time scale. This is particularly important for analyzing the aging process of transformers, as aging is a gradual process involving the cumulative effects of multiple factors over a long period. In this way, this layer can help identify potential risk points that may lead to equipment aging and provide a basis for subsequent risk assessment.

[0098] This multi-layered design with progressively increasing expansion coefficients enables TCN to effectively extract features across different time scales, thereby improving its ability to detect changes in transformer condition. Combining a 6-hour sliding window and a 30-minute sliding step, TCN can not only accurately identify short-term abnormal temperature fluctuations but also capture aging trends over longer time periods. This provides strong support for accurately calculating dynamic thresholds, thus enhancing the sensitivity and reliability of the entire aging detection system. Furthermore, due to TCN's high efficiency and lightweight nature, real-time monitoring and early warning can be achieved even in resource-constrained edge computing environments, greatly enhancing the system's practical value.

[0099] In some embodiments of this application, the dynamic threshold includes a temperature gradient amplitude threshold and a temperature standard deviation threshold. In step S150, referring to... Figure 3 The process of extracting temperature anomaly features from a three-dimensional temperature field mesh model based on a dynamic threshold includes, but is not limited to, the following steps.

[0100] Step S310: Calculate the temperature gradient magnitude of each grid cell and the temperature standard deviation of adjacent grid cells in the three-dimensional temperature field grid model in real time.

[0101] In step S310, based on the constructed three-dimensional temperature field grid model, the temperature gradient magnitude (i.e., the rate of temperature change in space) and the temperature standard deviation between that grid and its neighboring grids are calculated for each grid cell. This process quantifies the spatial non-uniformity and local fluctuations of temperature distribution, providing crucial data support for subsequent identification of abnormal areas. Real-time calculation ensures the dynamic nature and responsiveness of the aging state assessment, helping to promptly identify potential risk points.

[0102] Step S320: Mark grid cells whose temperature gradient magnitude exceeds the temperature gradient magnitude threshold as first-level abnormal regions.

[0103] In step S320, each grid cell is evaluated using a pre-set temperature gradient amplitude threshold. If the temperature gradient of a grid cell is significantly higher than the normal level, it indicates that there is a rapid temperature change in that area, which may be caused by local hotspots or material degradation. Marking these grid cells as first-level abnormal areas helps to quickly locate key parts that may have structural defects or early failures, providing a basis for subsequent risk level assessment.

[0104] Step S330: Mark grid cells whose temperature standard deviation exceeds the temperature standard deviation threshold as secondary anomaly regions.

[0105] In step S330, regions with large temperature fluctuations are identified by comparing the temperature standard deviation of each grid cell with its surrounding area to see if it exceeds a set dynamic standard deviation threshold. A large temperature standard deviation usually indicates unstable local heat distribution, which may be affected by aging, poor contact, or other non-uniform factors. Marking such regions as secondary anomaly regions can further refine the anomaly type, help distinguish the stability and randomness of local temperature rise, and improve the comprehensiveness and accuracy of the diagnosis.

[0106] Step S340: Extract the temperature gradient change rate of the primary anomaly region and the temperature fluctuation range of the secondary anomaly region as temperature anomaly features.

[0107] In step S340, based on the completion of the primary and secondary anomaly region division, representative temperature anomaly features are further extracted, including the temperature gradient change rate of the primary anomaly region (reflecting the severity of temperature changes in the anomaly region) and the temperature fluctuation range of the secondary anomaly region (reflecting the magnitude of temperature changes over time and space). These features not only enrich the data dimensions of aging analysis but also provide quantitative input parameters for subsequent input into the machine learning model for aging risk assessment, enhancing the scientific rigor and robustness of the model's judgment.

[0108] In some embodiments of this application, the machine learning models include support vector machine models and random forest models. These models are used to process temperature gradient features and insulating oil degradation feature data collected from the transformer to assess the aging degree of the transformer.

[0109] Support Vector Machines classify data by finding the optimal separating hyperplane, while Random Forests utilize multiple decision trees for voting decisions. Combining the two can improve the accuracy and robustness of predictions. Furthermore, transfer learning techniques are employed to enable the model to adapt to different types of transformers.

[0110] Machine learning models are adapted to different transformers through transfer learning. Transfer learning allows a model to apply knowledge learned on one or more source tasks to a target task, even if the data distribution of the target task differs from that of the source tasks. This method is particularly suitable for large equipment like transformers because different transformer models may have different structures and operating parameters, but their aging mechanisms share similarities. Through transfer learning, a trained model can be effectively adapted to new transformer models, reducing the amount of data and time required for retraining while maintaining high prediction accuracy.

[0111] In some embodiments of this application, the aging risk level includes low risk level, medium risk level, high risk level and emergency risk level.

[0112] When the aging risk level is low, it means the transformer's degree of polymerization is greater than 500, indicating that the insulation paper is close to new paper, the mechanical strength loss is less than 30%, and the aging rate is slow. Under this condition, the temperature gradient is not significantly abnormal, the furfural concentration in the oil is extremely low, the acid value of the insulating oil is normal, and chemical decay has not yet started.

[0113] When the aging risk level is medium, it means that the degree of polymerization of the transformer has decreased to below 500 but is still above 400. At this time, the insulation paper enters the accelerated aging period, with a mechanical strength loss between 30% and 50%. Special attention needs to be paid to the risk of local overheating. Typical characteristics include an increased local temperature gradient or excessive furfural concentration. In addition, the acid value may rise slightly, which will accelerate the cellulose degradation process.

[0114] When the aging risk level is high, it means the transformer's degree of polymerization has decreased to below 400, but is still above 250. At this point, the mechanical strength loss of the insulation paper exceeds 50%, indicating that its lifespan is nearing its end (a degree of polymerization of 250 is considered the end of its lifespan), and it is prone to breakdown under short-circuit impact. Multiple abnormal indicators show an accelerated aging rate. It is recommended to schedule a power outage for disassembly and maintenance within three months, replacing local insulation or performing oil filtration to slow further deterioration.

[0115] When the aging risk level is high, it means that the transformer's cohesion has decreased to below 250. At this point, the mechanical strength of the insulation paper is almost completely lost, and the transformer may collapse at any time due to electromagnetic force or thermal stress. In this situation, the transformer should be shut down immediately, the load transferred, the winding insulation replaced, and the emergency fault plan activated to prevent potential major accidents.

[0116] In some embodiments of this application, the process of calculating and outputting the remaining service life of a large transformer based on aging risk level, environmental compensation coefficient, and temperature anomaly characteristics, combined with the Weibull distribution model and the expected value method, includes, but is not limited to, the following steps.

[0117] Step S410: Based on the aging risk level, call the risk parameter mapping table to obtain the shape and scale parameters of the Weibull distribution model.

[0118] In step S410, based on the aging risk level output by the machine learning model, the corresponding Weibull distribution model parameters—namely, shape and scale parameters—are automatically matched using a pre-established risk parameter mapping table. Different aging levels represent differences in the aging degree of internal insulation materials, conductive components, etc., of the transformer, and the parameters of the Weibull distribution are closely related to its failure mode. This step provides a statistical model basis adapted to the current aging state for subsequent lifetime prediction, thereby improving the accuracy and personalization of remaining service life calculation.

[0119] Step S420: Based on the environmental compensation coefficient and temperature anomaly characteristics, combined with the Weibull distribution model and the expected value method, calculate and output the remaining service life of the large transformer.

[0120] In step S420, based on the obtained Weibull distribution parameters, this step further introduces the influence of environmental compensation coefficients and temperature anomaly characteristics on the equipment's operating status. It comprehensively considers the accelerating or inhibiting effects of the actual operating environment (such as temperature, humidity, wind speed, etc.) on the transformer's aging rate, as well as the additional thermal stress caused by local hot spots. By integrating the Weibull distribution using the expected value method, the expected time before the transformer fails or reaches the end of its lifespan, i.e., the remaining useful life (RUL), can be obtained. This method not only integrates statistical reliability theory but also incorporates real-time monitoring data, making the lifespan prediction results more practically applicable in engineering. It helps maintenance personnel develop scientific maintenance and replacement plans, improving the safety and economy of power system operation.

[0121] It should be noted that the Weibull distribution model is a probabilistic distribution model used to describe the lifespan characteristics of equipment or systems, and is particularly suitable for estimating the remaining life of electrical equipment with different aging modes. It characterizes the trend of failure rate changes over time through shape and scale parameters, and can flexibly adapt to various types of failure modes, such as early failures, random failures, and wear-out damage stages. Combined with the expected value method, which calculates the mean of the Weibull distribution to estimate the average remaining life of the equipment, it can provide scientific aging risk assessment and remaining life prediction for critical equipment such as large transformers. This method not only considers the inherent aging characteristics of the equipment but also incorporates the impact of the actual operating environment on lifespan, thereby improving the accuracy and reliability of the prediction results and contributing to the optimization of predictive maintenance strategies.

[0122] In some embodiments of this application, the remaining useful life satisfies the following formula (2):

[0123] (2);

[0124] In formula (2), Indicates the environmental compensation coefficient. Indicates the remaining service life. This represents the average rate of change of the temperature gradient within the temperature anomaly characteristics. Indicates the scale parameter. Indicates shape parameters, This represents the mean multiplier of the Weibull distribution model.

[0125] Formula (2) dynamically calculates the remaining service life of equipment by quantifying key temperature anomaly characteristics (such as the average rate of change of temperature gradient) during the equipment aging process, combined with the life characteristic parameters (scale and shape parameters) of the equipment materials and environmental compensation factors. This calculation mechanism incorporates the accelerating effect of temperature anomalies on equipment aging into the evaluation system and realizes the numerical transformation of life prediction through a statistical model (Weibull distribution mean multiplier). Its technical value lies in providing accurate time basis for equipment maintenance decisions. For example, in the scenario of large transformers, when an abnormal increase in temperature gradient is detected, the formula can shorten the estimated remaining life in real time, guiding maintenance personnel to formulate maintenance or replacement plans in advance, and avoiding unexpected downtime accidents caused by equipment aging failure.

[0126] In some embodiments of this application, assuming the aging risk level is high, the risk parameter mapping table is invoked to obtain the scale parameters of the Weibull distribution model. For 8000 hours, shape parameters It is 1.6; at this point, the mean rate of change of the temperature gradient is... The environmental compensation coefficient is 0.22. The mean multiplier is 1.35. Therefore, the mean multiplier of the Weibull distribution model is approximately 0.897, and the remaining useful life is calculated to be approximately 4142 hours.

[0127] In the same scenario, if we ignore the temperature gradient and environmental influences, that is... =0, With a value of 1, the mean multiplier of the Weibull distribution model is calculated to be approximately 0.897, resulting in a final calculated remaining service life of approximately 7176 hours. However, abnormal temperatures and high-temperature environments typically shorten the remaining service life of large transformers, highlighting the necessity of dynamic correction.

[0128] In some embodiments of this application, the process of marking abnormal areas of large transformers on a three-dimensional temperature field mesh model and issuing alarm information based on temperature anomaly characteristics, aging risk level, and remaining service life includes, but is not limited to, the following steps.

[0129] Step S510: Mark the abnormal regions corresponding to the temperature anomaly features in the three-dimensional temperature field mesh model, and send their corresponding coordinates to the remote monitoring platform.

[0130] In step S510, the previously analyzed temperature anomaly characteristics are first used to precisely mark the abnormal areas with potential problems in the three-dimensional temperature field mesh model. This method not only visually demonstrates which parts of the transformer may be at risk of aging or failure, but also provides clear location references for subsequent maintenance work. Furthermore, the specific coordinates of these marked areas are sent to a remote monitoring platform, enabling technicians not on-site to understand the equipment status in real time, conduct remote diagnosis and decision support, thus improving the efficiency and accuracy of fault handling.

[0131] Step S520: If the aging risk level is high risk level or emergency risk level, or the remaining service life is less than the preset service life threshold, send an alarm message to the remote monitoring platform.

[0132] In step S520, based on the aging risk level assessment results and remaining service life prediction, if the transformer is found to be at a high-risk or emergency risk level, or its remaining service life is below the preset safety threshold, the system will automatically trigger an alarm mechanism and send an alarm message to the remote monitoring platform. This step is crucial because it directly relates to the safe and stable operation of the power system. Timely alarms can prompt maintenance personnel to take swift action, such as arranging emergency repairs, replacing components, or adjusting the load, to avoid potential accidents and reduce economic losses and safety hazards. Furthermore, this data-driven risk warning mechanism is more efficient and accurate than traditional manual inspections, contributing to improved intelligent management of the entire power grid.

[0133] Secondly, refer to Figure 4 This application also provides a non-contact large transformer aging detection system, including a data acquisition module 610, a three-dimensional temperature field construction module 620, a temperature anomaly feature extraction module 630, and an aging risk prediction module 640.

[0134] The data acquisition module 610 is used to collect three-dimensional images, surface temperature distribution data and environmental data of large transformers in a non-contact manner, and to perform time alignment using GNSS synchronization timestamps.

[0135] The 3D temperature field construction module 620 is used to construct a 3D temperature field mesh model of a large transformer based on 3D images and surface temperature distribution data, and to perform mesh refinement processing on the winding area, core area and heat dissipation area.

[0136] The temperature anomaly feature extraction module 630 calculates the environmental compensation coefficient based on environmental data and historical environmental data. Based on the environmental compensation coefficient, surface temperature distribution data, and a preset time sliding window, a dynamic threshold is calculated using a prediction model. Temperature anomaly features are then extracted from the three-dimensional temperature field mesh model based on the dynamic threshold.

[0137] The aging risk prediction module 640 is used to input abnormal temperature features into a machine learning model and output the aging risk level of large transformers.

[0138] Furthermore, refer to Figure 5 This application provides a non-contact large transformer aging detection device, which is mobile and can be deployed on unmanned inspection equipment.

[0139] As a mobile, deployable intelligent monitoring device, this unit is integrated into unmanned inspection equipment and can flexibly move to different locations around transformers to achieve real-time, non-destructive testing of transformer operating status. The device eliminates the need for physical contact with the transformer, avoiding potential safety hazards and operational complexities associated with traditional testing methods, while simultaneously improving testing efficiency and adaptability. By deploying this device externally to the transformer, key parameters such as surface temperature distribution, environmental data, and 3D images can be continuously acquired and analyzed in real time, accurately assessing the transformer's aging degree and health status. This mobility makes the device suitable for various substation environments and different transformer models, enhancing the system's versatility and ease of on-site application, and providing strong support for intelligent inspection and condition-based maintenance in power systems.

[0140] The device includes a GNSS unit 710, a thermal imager 720, a 3D depth camera 730, an environmental sensor group 740, an edge computing unit 750, and a communication unit 760.

[0141] GNSS unit 710 provides synchronization timestamps to time-align surface temperature distribution data, 3D images, and environmental data. This unit provides precise time synchronization services, ensuring accurate alignment of the acquired surface temperature distribution data, 3D images, and environmental data along the timeline by generating synchronization timestamps. This is crucial for subsequent data analysis, as it guarantees the continuity and consistency of all collected information over time, thus providing a reliable time reference for non-contact aging detection of large transformers.

[0142] The 720 thermal imager is used to acquire surface temperature distribution data of large transformers. Precise measurement of the surface temperature of large transformers can identify potential hot spots or areas of abnormal temperature rise, which are often early signs of internal equipment failure or aging. Thermal imaging technology allows for rapid scanning of large areas without direct contact with the equipment, improving detection efficiency and safety.

[0143] The 3D depth camera 730 is used to acquire 3D images of large transformers. Based on these high-precision 3D images, a 3D model of the transformer can be constructed, which is of great significance for subsequent analysis of its physical structural changes and assessment of aging status. In addition, the 3D images also support mesh refinement processing for key components (such as winding areas and core areas) to improve the accuracy of local analysis.

[0144] The environmental sensor array 740 is used to collect environmental data from large transformers. Composed of multiple sensors, the array monitors environmental parameters around the transformer in real time, including temperature, humidity, and wind speed. This environmental data is crucial for understanding the impact of external conditions on transformer performance and forms the basis for calculating environmental compensation coefficients. It helps eliminate measurement biases caused by environmental factors and improves the accuracy of aging condition assessment.

[0145] The edge computing unit 750 includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned non-contact aging detection method for large transformers. The advantage of edge computing is that preliminary processing can be performed close to the data source, reducing latency and lowering the load on remote servers.

[0146] The communication unit 760 is used for communication with the remote monitoring platform. As a bridge connecting the field detection device and the remote monitoring platform, it ensures that the collected data and analysis results can be transmitted to the back-end management personnel in a timely manner. This not only facilitates real-time monitoring but also enables remote experts to quickly respond to any potential risk warnings and take necessary maintenance measures, greatly improving emergency response capabilities and management efficiency.

[0147] Furthermore, embodiments of this application provide a computer-readable storage medium storing a processor-executable program, which, when executed by a processor, is used to implement the aforementioned non-contact large transformer aging detection method.

[0148] In summary, the non-contact aging detection method, system, device, and medium for large transformers provided in this application have the following technical effects.

[0149] First, the embodiments of this application adopt a non-contact data acquisition method, including using a thermal imager to obtain surface temperature distribution data, a 3D depth camera to capture three-dimensional images, and an environmental sensor group to collect real-time environmental parameters, which avoids interference with the device's operating status and improves the security and efficiency of data acquisition.

[0150] Secondly, the embodiments of this application utilize the time synchronization service provided by the GNSS unit to ensure accurate alignment of all collected data in the time dimension, providing a reliable time reference for subsequent data analysis. The use of edge computing units not only enables preliminary processing near the data source, reducing data transmission latency, but also reduces the load on remote servers, improving the overall system's response speed and processing capacity.

[0151] Furthermore, this method, by constructing a three-dimensional temperature field grid model and combining it with environmental compensation coefficients and dynamic threshold calculations, can accurately identify the aging characteristics and abnormal areas of transformers. Employing machine learning models to predict aging risk levels and combining them with the Weibull distribution model and expected value method to assess remaining lifespan further enhances the accuracy and reliability of aging status assessment.

[0152] Furthermore, the device is designed to be mobile and deployed within unmanned inspection equipment, making it suitable for various substation environments and different transformer models, thus enhancing the system's versatility and ease of on-site application. The application of transfer learning technology allows the machine learning model to adapt to different transformer models, improving the system's flexibility and accuracy. The overall solution not only overcomes the limitations of traditional detection methods but also effectively reduces unnecessary maintenance costs and power outage time, providing strong support for intelligent inspection and condition-based maintenance in power systems, and possessing significant practical application value and socio-economic benefits.

[0153] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this application are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0154] Furthermore, although this application is described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding this application. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of ordinary skill of an engineer. Therefore, those skilled in the art can implement the application set forth in the claims using ordinary skill. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of this application, which is determined by the full scope of the appended claims and their equivalents.

[0155] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several programs to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0156] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequential list of executable programs for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, a program execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can retrieve and execute a program from or in conjunction with such a program execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit a program for use by or in conjunction with a program execution system, apparatus, or device.

[0157] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Additionally, computer-readable media can even be paper or other suitable media on which programs can be printed, for example, by optically scanning the paper or other media, then editing, interpreting, or, if necessary, processing it in a suitable manner to obtain the program electronically, and then storing it in computer memory.

[0158] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable program execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0159] In the foregoing description of this specification, the reference to terms such as "one embodiment / implementation," "another embodiment / implementation," or "certain embodiments / implementations," etc., indicates that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in an embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0160] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0161] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of the present invention.

Claims

1. A non-contact aging detection method for large transformers, characterized in that, Includes the following steps: The system acquires three-dimensional images, surface temperature distribution data, and environmental data of large transformers in a non-contact manner, and uses GNSS synchronization timestamps for time alignment. Based on the three-dimensional image and the surface temperature distribution data, a three-dimensional temperature field mesh model of the large transformer is constructed, and the mesh of the winding region, core region and heat dissipation region is refined. Based on the aforementioned environmental data and historical environmental data, the environmental compensation coefficient is calculated. Based on the environmental compensation coefficient, the surface temperature distribution data, and the preset time sliding window, a dynamic threshold is calculated using a prediction model. Based on the dynamic threshold, temperature anomaly features are extracted from the three-dimensional temperature field mesh model; The abnormal temperature features are input into a machine learning model, which outputs the aging risk level of the large transformer.

2. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The process of constructing a three-dimensional temperature field mesh model of the large transformer based on the three-dimensional image and the surface temperature distribution data, and refining the mesh for the winding region, core region, and heat dissipation region, includes the following steps: Based on the three-dimensional image, an initial mesh structure model of the large transformer is generated using a three-dimensional reconstruction algorithm; Using an image recognition algorithm, the winding region, the core region, and the heat dissipation region are identified and marked in the initial mesh structure model; Each data point in the surface temperature distribution data is matched and mapped to the initial mesh structure model to construct a three-dimensional temperature field mesh model of the large transformer, and local mesh refinement processing is performed on the winding region, core region and heat dissipation region.

3. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The environmental data includes real-time temperature, real-time humidity, and real-time wind speed, and the historical environmental data includes historical average temperature, historical average humidity, and historical average wind speed. The environmental compensation coefficient satisfies the following formula: ； in, This represents the environmental compensation coefficient; This indicates the real-time temperature. This indicates the real-time humidity. This indicates the real-time wind speed; This represents the historical average temperature. This represents the historical average humidity. This represents the historical average wind speed; This represents the temperature weighting coefficient. This represents the humidity weighting coefficient. This represents the wind speed weighting coefficient; .

4. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The dynamic thresholds include the temperature gradient amplitude threshold and the temperature standard deviation threshold; The step of extracting temperature anomaly features from the three-dimensional temperature field mesh model based on the dynamic threshold includes the following steps: The temperature gradient magnitude of each grid cell and the temperature standard deviation of adjacent grid cells in the three-dimensional temperature field grid model are calculated in real time. Grid cells whose temperature gradient amplitude exceeds the temperature gradient amplitude threshold are marked as first-level abnormal regions; Grid cells whose temperature standard deviation exceeds the temperature standard deviation threshold are marked as secondary anomaly regions; The temperature gradient change rate of the primary anomaly region and the temperature fluctuation range of the secondary anomaly region are extracted as the temperature anomaly features.

5. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The machine learning model includes a support vector machine model and a random forest model; the machine learning model adapts to different transformers through transfer learning.

6. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The time sliding window is 6 hours long and the sliding step is 30 minutes.

7. The non-contact aging detection method for large transformers according to claim 1, characterized in that, The prediction model includes a temporal convolutional network model; the temporal convolutional network model contains three dilated convolutional layers; wherein the dilation coefficient of the first dilated convolutional layer is 2, the dilation coefficient of the second dilated convolutional layer is 4, and the dilation coefficient of the third dilated convolutional layer is 8.

8. A non-contact aging detection system for large transformers, characterized in that, It includes a data acquisition module, a three-dimensional temperature field construction module, a temperature anomaly feature extraction module, and an aging risk prediction module; The data acquisition module is used to collect three-dimensional images, surface temperature distribution data and environmental data of large transformers in a non-contact manner, and to perform time alignment using GNSS synchronization timestamps; The three-dimensional temperature field construction module is used to construct a three-dimensional temperature field mesh model of the large transformer based on the three-dimensional image and the surface temperature distribution data, and to perform mesh refinement processing on the winding area, core area and heat dissipation area therein; The temperature anomaly feature extraction module is used to calculate an environmental compensation coefficient based on the environmental data and historical environmental data; calculate a dynamic threshold based on the environmental compensation coefficient, the surface temperature distribution data, and a preset time sliding window through a prediction model; and extract temperature anomaly features from the three-dimensional temperature field mesh model based on the dynamic threshold. The aging risk prediction module is used to input the abnormal temperature features into a machine learning model and output the aging risk level of the large transformer.

9. A non-contact aging detection device for large transformers, characterized in that, The device is mobile and deployed on unmanned inspection equipment, including a GNSS unit, a thermal imager, a 3D depth camera, an environmental sensor group, an edge computing unit, and a communication unit. The thermal imager is used to collect surface temperature distribution data of large transformers; The 3D depth camera is used to acquire three-dimensional images of the large transformer; The environmental sensor group is used to collect environmental data of the large transformer; The GNSS unit is used to provide a synchronization timestamp to time-align the surface temperature distribution data, the three-dimensional image, and the environmental data. The edge computing unit includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the non-contact aging detection method for large transformers as described in any one of claims 1 to 7. The communication unit is used to establish a communication connection with the remote monitoring platform.

10. A computer-readable storage medium storing a processor-executable program, characterized in that, The processor-executable program, when executed by the processor, is used to implement the non-contact aging detection method for large transformers as described in any one of claims 1 to 7.

Images