Transformer substation mutual inductor gap generation risk detection method, device, equipment and medium

Abstract

The invention discloses a transformer substation mutual inductor gap generation risk detection method, device, equipment and medium, and the method comprises the steps: building a gap detection model according to the spectral response difference of the reflectivity change spectrum of a base sealing ring of a transformer substation mutual inductor and the edge position of a metal contact surface under multispectral imaging under different elastic deformation characteristics; dividing a gap generation area according to the change speed of the reflectivity change spectrum, performing risk judgment, and performing gap detection training on the gap detection model according to the reflectivity characteristics of the leaked oil sample image and the ultraviolet fluorescence reaction of the leaked oil sample under ultraviolet irradiation. According to the method, the incidence relation between the reflectivity characteristics and the material characteristics is established, and the gap generation risk is detected by using the reflectivity change in the multispectral fusion image, the leakage oil distribution characteristics and the ultraviolet fluorescence reaction under the irradiation of the ultraviolet light source, so that the accuracy and efficiency of the state evaluation of the mutual inductor sealing ring are improved; and powerful support is provided for preventing sealing failure.

Description

Technical Field

[0001] The present invention relates to the technical field of transformer substation detection, and in particular to a transformer substation transformer gap generation risk detection method, device, equipment and medium. Background Art

[0002] In the supplementary inspection of substation blind spots, the detection of the transformer base sealing ring is a key link, especially the study of the spectral reflection law of the edge of the rubber ring and the metal contact surface.

[0003] Since the rubber ring will deform elastically during long-term operation, a tiny gap will be generated at the edge where the rubber ring contacts the metal. This gap will lead to a decrease in sealing performance, which will in turn cause oil leakage. In the prior art, it is often necessary to wait for obvious oil leakage to occur in the transformer before it can be detected through inspection technology. If the transformer with quality problems cannot be discovered and replaced in time during the supplementary inspection of the blind spot of the substation, it will threaten the safe operation of the substation.

[0004] Therefore, how to improve the detection accuracy of the transformer gap in the substation to avoid transformer oil leakage is a technical problem that needs to be solved urgently. Summary of the invention

[0005] The present invention provides a transformer substation transformer gap generation risk detection method, device, equipment and medium to achieve transformer substation transformer gap generation risk detection, thereby avoiding transformer oil leakage.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a method for detecting a transformer gap generation risk in a substation, comprising: Acquire first reflectivity characteristics of different regions at the edge of the base sealing ring and the metal contact surface of the transformer in the substation under multi-spectral imaging, and obtain a first reflectivity change map of the base sealing ring and the metal contact surface according to the first reflectivity characteristics of different regions; Building a gap detection model according to the spectral response differences of the first reflectivity variation map under different elastic deformation characteristics of the base sealing ring; According to the reflectivity characteristic change speed of the first reflectivity change map, a region with a change speed greater than a preset speed is divided into a gap generation region, a second reflectivity characteristic of the gap generation region under multispectral imaging is obtained, and a gap generation risk is pre-evaluated according to the second reflectivity characteristic to obtain a first risk detection result; Acquire a transformer oil leakage sample data set, and use an ultraviolet simulated light source to illuminate the oil leakage sample images in the oil leakage sample data set to obtain ultraviolet fluorescence reaction features; Perform gap detection training on the gap detection model based on the third reflectance feature, the ultraviolet fluorescence reaction feature, and the second reflectance feature of the leakage oil sample image in the leakage oil sample dataset under multispectral imaging; During the actual operation of the substation, input the image data of the instrument transformer collected in real time into the trained gap detection model to obtain the second risk detection result, and detect and evaluate the risk of gap generation of the target instrument transformer according to the first risk detection result and the second risk detection result.

[0007] Further, the method for obtaining the first reflectance feature of different regions at the edge position between the base seal ring and the metal contact surface of the substation instrument transformer under multispectral imaging, and obtaining the first reflectance change map of the base seal ring and the metal contact surface according to the first reflectance feature of different regions includes: Obtain the spectral images of different regions at the edge position between the base seal ring and the metal contact surface of the substation instrument transformer under multispectral imaging, and extract the first spectral reflectance data from the spectral images; Preprocess the first spectral reflectance data by using a Gaussian filter, and extract the first reflectance feature from the preprocessed first spectral reflectance data; Use the reflectance difference calculation method to calculate the difference between the first reflectance features at adjacent positions in the image, and obtain the first reflectance change map reflecting the change degree of the reflectance within a unit distance.

[0008] Further, the method for constructing a gap detection model according to the spectral response difference of the first reflectance change map under different elastic deformation characteristics of the base seal ring includes: Obtain the elastic deformation data sample set of the base seal ring when there is a gap in the instrument transformer; Extract several elastic deformation characteristics of the base seal ring when there is a gap in the instrument transformer according to the elastic deformation data sample set; Determine the mapping relationship between the elastic deformation characteristics and the gap generation conditions according to each elastic deformation characteristic and the corresponding first reflectance change map, and construct a gap detection model according to the mapping relationship.

[0009] Further, the method for dividing the region with a change speed greater than the preset speed into a gap generation region according to the change speed of the reflectance feature of the first reflectance change map, and obtaining the second reflectance feature of the gap generation region under multispectral imaging includes: Extract the reflectance jump data with a reflectance feature change speed greater than the preset speed in the first reflectance change map, and use the maximum inter-class variance method to segment the reflectance jump data to obtain the gap candidate region with a sudden change in reflectance features; Adjust the division boundaries of each of the gap candidate regions according to the regional area, edge orientation, and spectral intensity of the gap candidate regions to obtain gap generation regions, and acquire second reflectivity characteristics of the gap generation regions under multi-spectral imaging.

[0010] Further, pre-evaluate the gap generation risk based on the second reflectivity characteristics to obtain a first risk detection result, including: Calculate the change rate of the second reflectivity characteristics within the gap generation regions to obtain a second reflectivity change map within the gap generation regions; Screen the second reflectivity characteristics of the gap generation regions according to the second reflectivity change map. If there are points within the gap generation regions where the change rate value of the second reflectivity characteristics is greater than a preset reflectivity change threshold, it is determined that there is a gap generation risk in the gap generation regions corresponding to these points.

[0011] Further, the reflectivity change threshold is a multi-level reflectivity change threshold; The step of, if there are points within the gap generation regions where the change rate value of the second reflectivity characteristics is greater than a preset reflectivity change threshold, determining that there is a gap generation risk in the gap generation regions corresponding to these points, includes: If there are points within the gap generation regions where the change rate value of the second reflectivity characteristics is greater than a preset reflectivity change threshold, classify the gap generation risk level of the gap generation regions according to this change rate value and the multi-level reflectivity change threshold.

[0012] Further, acquire a mutual inductor oil leakage sample data set, irradiate the oil leakage sample images in the oil leakage sample data set with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; perform gap detection training on the gap detection model according to the third reflectivity characteristics, the ultraviolet fluorescence reaction characteristics, and the second reflectivity characteristics of the oil leakage sample images in the oil leakage sample data set, including: Acquire a mutual inductor oil leakage sample data set, and acquire the oil leakage spectrograms of each oil leakage sample image in the oil leakage sample data set under multi-spectral imaging; Extract the oil leakage distribution characteristics in the oil leakage spectrograms, and determine the oil leakage region boundaries in the oil leakage spectrograms according to the oil leakage distribution characteristics; Calculate the oil leakage diffusion speed data using the optical flow method based on at least two consecutive frames of oil leakage region boundaries, and establish an oil leakage diffusion trend prediction curve according to the oil leakage diffusion speed data; Acquire the surface image of the base seal ring corresponding to the oil leakage spectrogram, and use an edge detection algorithm to identify the elastic deformation information of the base seal ring in the surface image; Irradiate the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; Obtain the third reflectance characteristics of the leakage oil spectrogram under multispectral imaging. Based on the third reflectance characteristics, the ultraviolet fluorescence reaction characteristics, the second reflectance characteristics, the leakage oil diffusion trend prediction curve, and the elastic deformation information, perform gap detection training on the gap detection model.

[0013] Another embodiment of the present invention provides a substation transformer gap generation risk detection device, including: A feature extraction module, configured to obtain the first reflectance characteristics of different regions at the edge position between the base seal ring and the metal contact surface of the substation transformer under multispectral imaging, and obtain the first reflectance change map of the base seal ring and the metal contact surface according to the first reflectance characteristics of different regions; A model construction module, configured to construct a gap detection model according to the spectral response differences of the first reflectance change map under different elastic deformation characteristics of the base seal ring; A region division module, configured to divide the regions with a change speed greater than a preset speed into gap generation regions according to the change speed of the reflectance characteristics of the first reflectance change map, obtain the second reflectance characteristics of the gap generation regions under multispectral imaging, and pre-evaluate the gap generation risk according to the second reflectance characteristics to obtain a first risk detection result; An ultraviolet reaction module, configured to obtain a transformer leakage oil sample dataset, and irradiate the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; A model training module, configured to perform gap detection training on the gap detection model according to the third reflectance characteristics, the ultraviolet fluorescence reaction characteristics, and the second reflectance characteristics of the leakage oil sample images in the leakage oil sample dataset under multispectral imaging; A risk detection module, configured to input the real-time collected transformer image data into the trained gap detection model during the actual operation of the substation to obtain a second risk detection result, and detect and evaluate the gap generation risk of the target transformer according to the first risk detection result and the second risk detection result.

[0014] Further, the feature extraction module is used for: Obtain the spectral images of different regions at the edge position between the base seal ring and the metal contact surface of the substation transformer under multispectral imaging, and extract the first spectral reflectance data from the spectral images; Preprocess the first spectral reflectance data using a Gaussian filter, and extract first reflectance features from the preprocessed first spectral reflectance data; Use the reflectance difference calculation method to calculate the difference of the first reflectance features at adjacent positions in the image, and obtain a first reflectance change map reflecting the degree of change of the reflectance within a unit distance.

[0015] Further, the model construction module is used for: Obtain a sample set of elastic deformation data of the base sealing ring when there is a gap in the mutual inductor; Extract several elastic deformation features of the base sealing ring when there is a gap in the mutual inductor according to the elastic deformation data sample set; Determine the mapping relationship between the elastic deformation features and the gap generation conditions according to each elastic deformation feature and the corresponding first reflectance change map, and construct a gap detection model according to the mapping relationship.

[0016] Further, the region division module is used for: Extract reflectance jump data with a reflectance feature change speed greater than a preset speed in the first reflectance change map, and use the maximum inter-class variance method to segment the reflectance jump data to obtain a gap candidate region where the reflectance feature suddenly changes; Adjust the division boundary of each gap candidate region according to the region area, edge trend and spectral intensity of the gap candidate region to obtain a gap generation region and obtain the second reflectance feature of the gap generation region under multi-spectral imaging.

[0017] Further, the device further includes a risk judgment module: Calculate the change rate of the second reflectance feature in the gap generation region within the region to obtain a second reflectance change map within the gap generation region; Screen the second reflectance feature of the gap generation region according to the second reflectance change map. If there is a point where the change rate value of the second reflectance feature in the gap generation region is greater than a preset reflectance change threshold, it is determined that there is a risk of gap generation in the gap generation region corresponding to this point.

[0018] Further, the reflectance change threshold is a multi-level reflectance change threshold; The risk judgment module is further used for: The step that if there is a point where the change rate value of the second reflectance feature in the gap generation region is greater than a preset reflectance change threshold, it is determined that there is a risk of gap generation in the gap generation region corresponding to this point includes: If there is a point in the gap generation area where the change rate value of the second reflectivity feature is greater than the preset reflectivity change threshold, the gap generation risk level of the gap generation area is divided according to the change rate value and the multi-level reflectivity change threshold.

[0019] Further, the model training module is used for: Obtain a dataset of mutual inductor oil leakage samples, and obtain the oil leakage spectrograms of each oil leakage sample image in the dataset of oil leakage samples under multi-spectral imaging; Extract the oil leakage distribution features in the oil leakage spectrogram, and determine the oil leakage area boundary in the oil leakage spectrogram according to the oil leakage distribution features; Calculate the oil leakage diffusion speed data by using the optical flow method according to at least two consecutive frames of oil leakage area boundaries, and establish an oil leakage diffusion trend prediction curve according to the oil leakage diffusion speed data; Obtain the surface image of the base sealing ring corresponding to the oil leakage spectrogram, and identify the elastic deformation information of the base sealing ring in the surface image by using an edge detection algorithm; Irradiate the oil leakage sample image in the dataset of oil leakage samples with an ultraviolet simulation light source to obtain an ultraviolet fluorescence reaction feature; Obtain the third reflectivity feature of the oil leakage spectrogram under multi-spectral imaging, and perform gap detection training on the gap detection model according to the third reflectivity feature, the ultraviolet fluorescence reaction feature, the second reflectivity feature, the oil leakage diffusion trend prediction curve, and the elastic deformation information.

[0020] Another embodiment of the present invention provides a computer device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, the method for detecting the risk of gap generation in a substation mutual inductor as described above is implemented.

[0021] Another embodiment of the present invention provides a computer-readable storage medium, which stores a computer program. When the device where the computer-readable storage medium is located executes the computer program, the method for detecting the risk of gap generation in a substation mutual inductor as described above is implemented.

[0022] Compared with the prior art, the beneficial effects of the embodiments of the present invention are at least one of the following: By acquiring the multi-spectral image data of the edge of the sealing ring of the mutual inductor base and the metal contact surface, analyzing the spectral reflection law of the rubber ring and the metal contact surface, and establishing the correlation between the reflectivity characteristics and the material characteristics; using these data to construct a gap detection model, and by analyzing the reflectivity changes and the leakage oil distribution characteristics in the multi-spectral fusion image, the accurate evaluation of the state of the base sealing ring is realized; the present invention integrates the multi-spectral data in the visible light, infrared light and ultraviolet light bands, classifies and processes the image features, and finally outputs the judgment result of the gap generation risk level, improving the accuracy and efficiency of the state evaluation of the mutual inductor sealing ring and providing strong support for preventing seal failure. Description of the Drawings

[0023] Figure 1 It is a flowchart of the steps of the method for detecting the risk of gap generation of the substation mutual inductor provided by the embodiment of the present invention; Figure 2 It is a structural block diagram of the device for detecting the risk of gap generation of the substation mutual inductor provided by the embodiment of the present invention; Figure 3 It is a structural diagram of the computer device provided by the embodiment of the present invention. Detailed Embodiments

[0024] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0025] In the description of the present application, the terms "first", "second", "third", etc. are only used for descriptive purposes, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, the features defined with "first", "second", "third", etc. may explicitly or implicitly include one or more of such features. In the description of the present application, unless otherwise specified, the meaning of "a plurality" is two or more.

[0026] In the description of the present application, it should be noted that unless otherwise clearly defined and limited, the terms "installation", "connection", and "coupling" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected or indirectly connected through an intermediate medium, and it can be the communication inside two components. The terms "vertical", "horizontal", "left", "right", "up", "down" and similar expressions used herein are only for the purpose of illustration, rather than indicating or implying that the device or component referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus cannot be construed as a limitation on the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0027] In the description of the present application, it should be noted that unless otherwise defined, all technical and scientific terms used in the present invention have the same meaning as commonly understood by those skilled in the technical field to which this invention belongs. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments, and are not intended to limit the present invention. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to specific circumstances.

[0028] An embodiment of the present invention provides a method for detecting the risk of substation transformer gap generation. Specifically, please refer to Figure 1 , Figure 1 which shows the step flowchart of the method for detecting the risk of substation transformer gap generation in one of the embodiments of the present invention, including steps S11 to S16: Step S11: Obtain the first reflectivity characteristics of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer under multispectral imaging, and obtain the first reflectivity change map of the base seal ring and the metal contact surface according to the first reflectivity characteristics of different regions.

[0029] Multispectral imaging technology can capture spectral information in different bands. These spectral information reflect the reflection, transmission, and absorption characteristics of substances at different wavelengths. For power equipment such as substation transformers, the material, state (such as oxidation, corrosion, pollution, etc.) and contact situation of the base seal ring and the metal contact surface will be reflected in the spectral image.

[0030] Due to the different spectral reflection characteristics at the edge of the contact surface between the rubber ring and the metal, when a gap occurs, the reflected spectrum will change, and this change is manifested as a difference in the reflection intensity of a local area in multi-spectral imaging. However, due to the material differences between the rubber ring and the metal surface, the spectral reflection law is complex, and multi-spectral fusion can synergistically process spectral data in different bands to extract the reflection characteristics at the edge of the contact surface between the rubber ring and the metal. When the elastic deformation of the sealing ring causes a gap to occur, the spectral reflection characteristics of the gap area will be significantly different from those of the normal contact area.

[0031] Obtain the spectral image under multi-spectral imaging at the edge position of the contact surface between the base sealing ring and the metal of the substation transformer. Extract the first spectral reflectance data from the spectral image. Compared with traditional visible light imaging, multi-spectral imaging can provide richer information, which helps to more accurately identify and analyze the state of the equipment. Multi-spectral imaging is not affected by lighting conditions and can work stably in various environments, improving the reliability and applicability of detection. Spectral reflectance data is the data that describes the ratio of the reflected light intensity to the incident light intensity of a substance at different wavelengths, and it directly reflects the reflection characteristics of the substance.

[0032] Preprocess the first spectral reflectance data using a Gaussian filter, which can remove noise and outliers in the data, making the data smoother and more reliable. Extract the first reflectance feature from the preprocessed first spectral reflectance data.

[0033] The reflectance difference calculation method reflects the change degree of the reflectance within a unit distance by calculating the difference in reflectance at adjacent positions. Use the reflectance difference calculation method to calculate the difference in the first reflectance features at adjacent positions in the image, and obtain the first reflectance change map that reflects the change degree of the reflectance within a unit distance, which can intuitively reflect the spatial change of the reflectance.

[0034] Step S12, construct a gap detection model based on the spectral response differences of the first reflectance change map under different elastic deformation characteristics of the base sealing ring.

[0035] In the detection and maintenance of transformers (especially current transformers), ensuring their sealing performance is crucial. The sealing performance directly affects the operation stability and service life of the transformer. And the base sealing ring, as a key part of the transformer sealing structure, its elastic deformation can indirectly reflect the sealing state of the transformer and whether there is a gap. Therefore, it is necessary to conduct in-depth research on the elastic deformation of the base sealing ring to construct an effective gap detection model.

[0036] Obtain the elastic deformation data sample set of the base sealing ring when there is a gap in the transformer, and extract several elastic deformation characteristics of the base sealing ring when there is a gap in the transformer according to the elastic deformation data sample set.

[0037] According to the elastic deformation characteristics and the corresponding first reflectivity change map, the mapping relationship between the elastic deformation characteristics and the gap generation conditions is determined, and a gap detection model is constructed based on the mapping relationship, so as to realize the quantitative evaluation of the sealing performance of the mutual inductor and improve the accuracy and reliability of the detection.

[0038] Step S13: Divide the region where the change speed of the reflectivity characteristics in the first reflectivity change map is greater than the preset speed into a gap generation region, obtain the second reflectivity characteristics of the gap generation region under multi-spectral imaging, and pre-evaluate the gap generation risk according to the second reflectivity characteristics to obtain a first risk detection result.

[0039] After obtaining the second reflectivity characteristics of the region where the reflectivity change speed is greater than the preset speed, the gap generation risk can be pre-evaluated according to the second reflectivity characteristics to obtain a first risk detection result. In actual applications, the gap generation risk can be estimated through this preliminary detection step. The specific process is as follows: Since under normal circumstances, the difference in reflectivity between adjacent regions will be maintained within a small range, when the reflectivity characteristics change suddenly, it means that a gap may appear in the nearby region. The specific judgment method is as follows: Extract the reflectivity jump data from the first reflectivity change map, and use the maximum between-class variance method to segment the reflectivity jump data, so as to initially identify the region where the reflectivity characteristics change greatly and obtain the gap candidate region where the reflectivity characteristics change suddenly.

[0040] Adjust the division boundary of each gap candidate region according to the region area, edge trend and spectral intensity of the gap candidate region to obtain the gap generation region and obtain the second reflectivity characteristics of the gap generation region under multi-spectral imaging. After obtaining the gap generation region, further obtain the second reflectivity characteristics of these regions under multi-spectral imaging. The second reflectivity characteristic value is used to further screen out the gap generation risk area from the gap candidate region.

[0041] Obtain the second reflectivity change map of the gap generation region according to the change trend of the continuous points in the gap generation region of the second reflectivity characteristics; if there are points in the second reflectivity change map where the second reflectivity change value is greater than the reflectivity change threshold, it is determined that the corresponding gap generation region has a gap generation risk.

[0042] After obtaining the second reflectivity feature of the gap candidate region, it is necessary to further accurately judge this region. Specifically: calculate the change rate of the second reflectivity feature in the gap generation region within the region to obtain the second reflectivity change map in the gap generation region. It is possible to promptly discover regions where the reflectivity feature may change significantly due to gap generation based on the second reflectivity change map, thereby identifying the gap generation risk.

[0043] Screen the second reflectivity feature of the gap generation region according to the second reflectivity change map. Divide multiple levels of reflectivity change thresholds according to different gap generation risk levels. If there is one or more points in the gap generation region whose corresponding reflectivity change values in the second reflectivity change map are greater than the reflectivity change threshold, it is determined that there is a gap generation risk in the corresponding gap generation region, and the corresponding gap generation risk level is determined according to the actual reflectivity change value and the corresponding multiple levels of reflectivity change thresholds.

[0044] Through this method, the gap generation risk can be analyzed and detected quickly, which is applicable to the case of simple scenarios, helps the inspectors to initially screen the current transformers and regions where gaps may occur, and is helpful for the inspectors to quickly master the situation in practical applications.

[0045] This second reflectivity feature is also used in subsequent processing steps to perform further more accurate refined modeling detection in combination with a series of data such as infrared reaction data.

[0046] The multiple levels of reflectivity change thresholds are different levels of thresholds set according to experimental data, experience, or industry standards. These thresholds reflect the possible change range of the reflectivity feature under different risk levels. Through the multiple levels of reflectivity change thresholds, the risk level of the gap generation region can be divided more meticulously, thereby providing a more accurate basis for subsequent processing measures, realizing the fine division and differential division of the gap generation risk, and providing a basis for subsequent gap generation risk control.

[0047] Step S14, obtain the transformer oil leakage sample data set, and irradiate the oil leakage sample images in the oil leakage sample data set with an ultraviolet simulation light source to obtain the ultraviolet fluorescence reaction feature.

[0048] To train the gap detection model, a large amount of oil leakage sample data is required as input. Obtain the transformer oil leakage sample data set, and these data sets contain various oil leakage situations to ensure the generalization ability of the model.

[0049] Obtain the leakage oil spectrograms of each leakage oil sample image in the leakage oil sample dataset under multi-spectral imaging, extract the leakage oil distribution characteristics in the leakage oil spectrograms. The leakage oil will exhibit specific reflectance characteristics under multi-spectral imaging, and these characteristics can help identify the distribution area of the leakage oil. Determine the boundary of the leakage oil area in the leakage oil spectrogram according to the leakage oil distribution characteristics. Defining the boundary of the leakage oil area helps to more accurately calculate the diffusion rate of the leakage oil and predict its diffusion trend.

[0050] Irradiate the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics. Insulating oils such as transformer oil will produce fluorescence reactions under ultraviolet irradiation, making the leakage oil more obvious in the image, forming a sharp contrast with the surrounding background for easy detection.

[0051] Understanding the diffusion rate of the leakage oil is crucial for assessing its potential risks and formulating emergency treatment measures. Therefore, in this embodiment, the diffusion rate data of the leakage oil is calculated using the optical flow method based on at least two frames of leakage oil boundaries, and a leakage oil diffusion trend prediction curve is established according to the leakage oil diffusion rate data to intuitively understand the diffusion trend and potential risks of the leakage oil.

[0052] Obtain the surface image of the base seal ring corresponding to the leakage oil spectrogram, and use an edge detection algorithm to identify the elastic deformation information of the base seal ring in the surface image.

[0053] Step S15: Perform gap detection training on the gap detection model according to the third reflectance characteristic, the ultraviolet fluorescence reaction characteristic, and the second reflectance characteristic of the leakage oil sample image in the leakage oil sample dataset under multi-spectral imaging.

[0054] Obtain the third reflectance characteristic of the leakage oil spectrogram under multi-spectral imaging, and perform gap detection training on the gap detection model according to the third reflectance characteristic, the ultraviolet fluorescence reaction characteristic, the second reflectance characteristic, the leakage oil diffusion trend prediction curve, and the elastic deformation information.

[0055] Clean and preprocess the data of the third reflectance characteristic, the ultraviolet fluorescence reaction characteristic, the leakage oil diffusion trend prediction curve, and the elastic deformation information, removing the abnormal values and missing values where the data has jumps to ensure the quality of the data.

[0056] Divide the processed data into a training set, a validation set, and a test set. During the division process, it is necessary to ensure that the samples in each dataset are representative and evenly distributed.

[0057] Preferably, in order to improve the local perception of the model during the gap detection process, the base model used in this embodiment to construct the gap detection model is a convolutional neural network model. The processed third reflectivity feature, ultraviolet fluorescence reaction feature, leakage oil diffusion trend prediction curve, and elastic deformation information are fused to obtain a high-dimensional feature vector as the input of the gap detection model. During the fusion process, weight distribution can be performed on the third reflectivity feature, ultraviolet fluorescence reaction feature, leakage oil diffusion trend prediction curve, and elastic deformation information according to the importance of each feature and the actual situation using feature selection methods, such as feature selection methods based on information gain, etc., which will not be elaborated here.

[0058] The mean square error function is used as the loss function of the gap detection model, and the average value of the squared difference between the predicted value and the true value of the gap detection model is calculated based on the output of each round of the model.

[0059] During the training process, the stochastic gradient descent method is used to calculate the gradient of the loss function with respect to the parameters of the gap detection model to minimize the loss function, and the parameters of the model are updated through backpropagation.

[0060] Repeat the above steps until the loss value of the gap detection model on the training set converges to a value below the preset deviation value, which means that the deviation between the detection result of the gap in the image by the gap detection model and the true result is less than the acceptable deviation range, indicating that the training effect meets the standard. End the model training, and use the model at this time as the model after the optimized training to perform subsequent detection tasks.

[0061] Preferably, during the training process, at the end of each round of iteration, the model is verified using the validation set, and the above-mentioned preset deviation value is updated according to the detection accuracy at the end of each round. Specifically, the verification metrics include accuracy, F1 score, etc.

[0062] The introduction of the leakage oil sample data set in this step is to obtain the third reflectivity features of the base sealing rings corresponding to different degrees of leakage oil from slight to severe, so as to predict the reflectivity features of the base sealing rings that are about to have gaps and thus cause oil leakage phenomena, so as to detect problematic transformers during the period when there is a risk of gap generation and avoid the occurrence of oil leakage phenomena.

[0063] The ultraviolet fluorescence reaction can excite the fluorescent substances in the leakage oil, making them emit visible light under ultraviolet irradiation. Even trace amounts of leakage oil can be detected, significantly improving the detection sensitivity. Detecting leakage oil through the ultraviolet fluorescence reaction can further verify whether there is a leakage risk in the gap generation area and avoid misjudgment or missed judgment. Combining multi-spectral imaging and ultraviolet fluorescence reaction can simultaneously obtain data on the risk of gap generation and leakage oil risk, providing a more reliable basis for the detection results.

[0064] Step S16, during the actual operation of the substation, input the image data of the instrument transformer collected in real time into the trained gap detection model to obtain a second risk detection result, and detect and evaluate the risk of gap generation of the target instrument transformer according to the first risk detection result and the second risk detection result.

[0065] Among them, the image data of the instrument transformer obtained by real-time collection includes the image data of different regions at the edge position of the base sealing ring and the metal contact surface of the instrument transformer in the target substation under various directions and lighting conditions.

[0066] Input the collected image data of the instrument transformer into the gap detection model for detection to obtain a more accurate and refined second risk detection result.

[0067] According to the preliminary first risk detection result pre-evaluated in step S13 and the second risk detection result generated by the gap detection model, jointly realize the detection and evaluation of the risk of gap generation of the target instrument transformer. The method for detecting the risk of gap generation of the substation instrument transformer of the present invention obtains multi-spectral image data of the edge of the base sealing ring and the metal contact surface of the instrument transformer, analyzes the spectral reflection law of the rubber ring and the metal contact surface, and establishes the correlation between the reflectivity characteristics and the material characteristics; uses these data to construct a gap detection model, and realizes the accurate evaluation of the state of the base sealing ring by analyzing the reflectivity change and the leakage oil distribution characteristics in the multi-spectral fusion image; the present invention integrates multi-spectral data in the visible light, infrared light and ultraviolet light bands, classifies and processes the image features, and finally outputs the judgment result of the gap generation risk level, improving the accuracy and efficiency of the state evaluation of the instrument transformer sealing ring, and providing strong support for preventing seal failure.

[0068] The embodiment of the present invention also provides a device for detecting the risk of gap generation of a substation instrument transformer, which is used to execute the method for detecting the risk of gap generation of a substation instrument transformer as described above. Figure 2 It is a structural block diagram of the device for detecting the risk of gap generation of the substation instrument transformer according to the embodiment of the present invention. The device includes: A feature extraction module 21, configured to obtain the first reflectivity characteristics of different regions at the edge position of the base sealing ring and the metal contact surface of the substation instrument transformer under multi-spectral imaging, and obtain the first reflectivity change map of the base sealing ring and the metal contact surface according to the first reflectivity characteristics of different regions. A model construction module 22, configured to construct a gap detection model according to the spectral response difference of the first reflectivity change map under different elastic deformation characteristics of the base sealing ring. The region division module 23 is configured to divide, according to the change speed of the reflectivity feature of the first reflectivity change map, the region where the change speed is greater than the preset speed into a gap generation region, obtain the second reflectivity feature of the gap generation region under multispectral imaging, pre-evaluate the gap generation risk according to the second reflectivity feature, and obtain a first risk detection result; The ultraviolet reaction module 24 is configured to obtain a dataset of transformer oil leakage samples, and irradiate the oil leakage sample images in the dataset of transformer oil leakage samples with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; The model training module 25 is configured to perform gap detection training on the gap detection model according to the third reflectivity feature of the oil leakage sample images in the dataset of transformer oil leakage samples under multispectral imaging, the ultraviolet fluorescence reaction characteristics, and the second reflectivity feature; The risk detection module 26 is configured to, during the actual operation of the substation, input the transformer image data collected in real time into the trained gap detection model to obtain a second risk detection result, and detect and evaluate the gap generation risk of the target transformer according to the first risk detection result and the second risk detection result.

[0069] Among them, the feature extraction module is used for: Obtain spectral images of different regions at the edge position of the base sealing ring and the metal contact surface of the substation transformer under multispectral imaging, and extract first spectral reflectivity data from the spectral images; Preprocess the first spectral reflectivity data by using a Gaussian filter, and extract first reflectivity features from the preprocessed first spectral reflectivity data; Use the reflectivity difference calculation method to calculate the difference of the first reflectivity features at adjacent positions in the image, and obtain a first reflectivity change map reflecting the change degree of the reflectivity within a unit distance.

[0070] The model construction module is used for: Obtain a sample set of elastic deformation data of the base sealing ring when there is a gap in the transformer; Extract several elastic deformation features of the base sealing ring when there is a gap in the transformer according to the elastic deformation data sample set; Determine the mapping relationship between the elastic deformation features and the gap generation conditions according to each elastic deformation feature and the corresponding first reflectivity change map, and construct a gap detection model according to the mapping relationship.

[0071] The region division module is used for: Extract the reflectivity jump data with a reflectivity feature change speed greater than a preset speed from the first reflectivity change map, and use the Otsu method to segment the reflectivity jump data to obtain candidate regions for gaps where the reflectivity features change abruptly; Adjust the division boundaries of the candidate regions for gaps according to the area, edge orientation, and spectral intensity of the candidate regions for gaps to obtain gap generation regions and acquire the second reflectivity features of the gap generation regions under multispectral imaging.

[0072] The device further includes a risk judgment module: Calculate the change rate of the second reflectivity features within the gap generation regions to obtain the second reflectivity change map within the gap generation regions; Screen the second reflectivity features of the gap generation regions according to the second reflectivity change map. If there are points within the gap generation regions where the change rate value of the second reflectivity features is greater than a preset reflectivity change threshold, it is determined that there is a risk of gap generation in the gap generation regions corresponding to these points.

[0073] The reflectivity change threshold is a multi-level reflectivity change threshold; The risk judgment module is further used for: The step of determining that there is a risk of gap generation in the gap generation regions corresponding to the points if there are points within the gap generation regions where the change rate value of the second reflectivity features is greater than a preset reflectivity change threshold includes: If there are points within the gap generation regions where the change rate value of the second reflectivity features is greater than a preset reflectivity change threshold, classify the risk of gap generation in the gap generation regions according to the change rate value and the multi-level reflectivity change threshold.

[0074] The model training module is used for: Obtain a dataset of current transformer oil leakage samples, and acquire the oil leakage spectrograms of each oil leakage sample image in the oil leakage sample dataset under multispectral imaging; Extract the oil leakage distribution features in the oil leakage spectrograms, and determine the oil leakage region boundaries in the oil leakage spectrograms according to the oil leakage distribution features; Calculate the oil leakage diffusion speed data using the optical flow method based on the oil leakage region boundaries of at least two consecutive frames, and establish an oil leakage diffusion trend prediction curve according to the oil leakage diffusion speed data; Obtain the surface image of the base seal ring corresponding to the oil leakage spectrogram, and use an edge detection algorithm to identify the elastic deformation information of the base seal ring in the surface image; Irradiate the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; Obtain the third reflectance characteristics of the leakage oil spectrogram under multispectral imaging. Based on the third reflectance characteristics, the ultraviolet fluorescence reaction characteristics, the second reflectance characteristics, the leakage oil diffusion trend prediction curve, and the elastic deformation information, perform gap detection training on the gap detection model.

[0075] The technical features and technical effects of the device proposed in the embodiments of the present invention are the same as those of the method proposed in the embodiments of the present invention, and will not be elaborated here. Each module in the above device can be implemented in whole or in part by software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above modules.

[0076] The embodiments of the present invention also provide a computer-readable storage medium, which includes a stored computer program; wherein, when the computer program runs, it controls the device where the computer-readable storage medium is located to execute the substation transformer gap generation risk detection method as described above.

[0077] The embodiments of the present invention also provide a computer device, Figure 3 which is a structural block diagram of a preferred embodiment of a computer device provided by the present invention. The computer device includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the substation transformer gap generation risk detection method as described above.

[0078] Preferably, the computer program can be divided into one or more modules / units (such as computer program 1, computer program 2,...). The one or more modules / units are stored in the memory and executed by the processor to complete the present invention. The one or more modules / units can be a series of computer program instruction segments capable of completing specific functions, and these instruction segments are used to describe the execution process of the computer program in the computer device.

[0079] The processor may be a Central Processing Unit (CPU), or may also be other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor, or the processor may also be any conventional processor. The processor is the control center of the computer device and connects various parts of the computer device through various interfaces and lines.

[0080] The memory mainly includes a program storage area and a data storage area. Among them, the program storage area can store an operating system, application programs required for at least one function, etc., and the data storage area can store relevant data, etc. In addition, the memory may be a high-speed random access memory, or may also be a non-volatile memory, such as a plug-in hard disk, a SmartMedia Card (SMC), a Secure Digital (SD) card, a Flash Card, etc., or the memory may also be other volatile solid-state storage devices.

[0081] It should be noted that the above computer device may include, but is not limited to, a processor and a memory. Those skilled in the art can understand that Figure 3 The structural block diagram is only an example of the computer device and does not constitute a limitation on the computer device. It may include more or fewer components than those shown in the figure, or combine certain components, or different components.

[0082] In summary, compared with the prior art, the beneficial effects of the substation transformer mutual inductor gap generation risk detection method, device, equipment and medium provided by the embodiments of the present invention are at least one of the following: By acquiring the multi-spectral image data of the edge of the metal contact surface of the transformer mutual inductor base sealing ring, analyzing the spectral reflection law of the rubber ring and the metal contact surface, and establishing the correlation relationship between the reflectivity characteristics and the material characteristics; using these data to construct a gap detection model, and by analyzing the reflectivity change and the leakage oil distribution characteristics in the multi-spectral fusion image, the accurate evaluation of the state of the base sealing ring is realized; the present invention integrates the multi-spectral data in the visible light, infrared light and ultraviolet light bands, classifies and processes the image features, and finally outputs the judgment result of the gap generation risk level, improving the accuracy and efficiency of the evaluation of the state of the transformer mutual inductor sealing ring, and providing strong support for preventing seal failure.

[0083] The above-described embodiments merely represent several implementation manners of the present invention. The description thereof is relatively specific and detailed, but it should not be construed as a limitation to the scope of the patent for the present invention. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention. Therefore, the protection scope of the patent for the present invention shall be subject to the appended claims.

Claims

1. A method for detecting the risk of generating gaps in transformers of a substation, characterized in that, Including: Obtain the first reflectivity characteristics of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer in multispectral imaging, and obtain the first reflectivity change map of the base seal ring and the metal contact surface according to the first reflectivity characteristics of different regions; Construct a gap detection model according to the spectral response differences of the first reflectivity change map under different elastic deformation characteristics of the base seal ring; Divide the regions with a change speed greater than a preset speed into gap generation regions according to the change speed of the reflectivity characteristics of the first reflectivity change map, obtain the second reflectivity characteristics of the gap generation regions in multispectral imaging, and pre-evaluate the gap generation risk according to the second reflectivity characteristics to obtain the first risk detection result; Obtain the transformer oil leakage sample data set, and irradiate the transformer oil leakage sample images in the leakage oil sample data set with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; Perform gap detection training on the gap detection model according to the third reflectivity characteristics, the ultraviolet fluorescence reaction characteristics, and the second reflectivity characteristics of the transformer oil leakage sample images in the leakage oil sample data set in multispectral imaging; During the actual operation of the substation, input the real-time collected transformer image data into the trained gap detection model to obtain the second risk detection result, and detect and evaluate the gap generation risk of the target transformer according to the first risk detection result and the second risk detection result.

2. The risk detection method for the transformer gap in a substation according to claim 1, characterized in that The obtaining of the first reflectivity characteristics of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer, and obtaining the first reflectivity change map of the base seal ring and the metal contact surface according to the first reflectivity characteristics of different regions includes: Obtain the spectral images of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer in multispectral imaging, and extract the first spectral reflectivity data from the spectral images; Preprocess the first spectral reflectivity data by using a Gaussian filter, and extract the first reflectivity characteristics from the preprocessed first spectral reflectivity data; Use the reflectivity difference calculation method to calculate the difference of the first reflectivity characteristics at adjacent positions in the image, and obtain the first reflectivity change map reflecting the change degree of the reflectivity within a unit distance.

3. The risk detection method for the substation transformer mutual inductor gap generation as described in claim 1, characterized in that, The constructing of the gap detection model according to the spectral response differences of the first reflectivity change map under different elastic deformation characteristics of the base seal ring includes: Obtain the elastic deformation data sample set of the base seal ring when there is a gap in the transformer; Extract several elastic deformation characteristics of the base seal ring when there is a gap in the transformer according to the elastic deformation data sample set; Determine the mapping relationship between the elastic deformation characteristics and the gap generation conditions according to each elastic deformation characteristic and the corresponding first reflectivity change map, and construct a gap detection model according to the mapping relationship.

4. The risk detection method for the substation transformer mutual inductor gap generation according to claim 1, characterized in that, Dividing the region where the change speed of the reflectivity feature according to the first reflectivity change map is greater than the preset speed into a gap generation region, and obtaining the second reflectivity feature of the gap generation region under multispectral imaging, includes: Extracting the reflectivity jump data with a change speed of the reflectivity feature greater than the preset speed in the first reflectivity change map, and segmenting the reflectivity jump data by the maximum inter-class variance method to obtain the gap candidate regions where the reflectivity feature changes suddenly; Adjusting the division boundaries of the gap candidate regions according to the region area, edge trend and spectral intensity of the gap candidate regions to obtain the gap generation regions and obtaining the second reflectivity feature of the gap generation regions under multispectral imaging.

5. The risk detection method for the substation transformer mutual inductor gap generation as described in claim 1, characterized in that, Pre-evaluating the gap generation risk according to the second reflectivity feature to obtain the first risk detection result, includes: Calculating the change rate of the second reflectivity feature in the gap generation region within the region to obtain the second reflectivity change map in the gap generation region; Screening the second reflectivity feature of the gap generation region according to the second reflectivity change map. If there is a point where the change rate value of the second reflectivity feature in the gap generation region is greater than the preset reflectivity change threshold, it is determined that there is a gap generation risk in the gap generation region corresponding to this point.

6. The risk detection method for the substation transformer mutual inductor gap generation according to claim 5, characterized in that, The reflectivity change threshold is a multi-level reflectivity change threshold; The step of, if there is a point where the change rate value of the second reflectivity feature in the gap generation region is greater than the preset reflectivity change threshold, determining that there is a gap generation risk in the gap generation region corresponding to this point, includes: If there is a point where the change rate value of the second reflectivity feature in the gap generation region is greater than the preset reflectivity change threshold, dividing the gap generation risk level of the gap generation region according to the change rate value and the multi-level reflectivity change threshold.

7. The risk detection method for the substation transformer mutual inductor gap generation as described in claim 1, wherein, Obtaining the transformer oil leakage sample data set, irradiating the oil leakage sample images in the oil leakage sample data set with an ultraviolet simulation light source to obtain the ultraviolet fluorescence reaction characteristics; training the gap detection model according to the third reflectivity feature, the ultraviolet fluorescence reaction characteristics and the second reflectivity feature of the oil leakage sample images in the oil leakage sample data set under multispectral imaging, includes: Obtaining the transformer oil leakage sample data set, and obtaining the oil leakage spectrograms of the oil leakage sample images in the oil leakage sample data set under multispectral imaging; Extracting the oil leakage distribution characteristics in the oil leakage spectrogram, and determining the oil leakage region boundary in the oil leakage spectrogram according to the oil leakage distribution characteristics; Calculating the oil leakage diffusion speed data by the optical flow method according to at least two consecutive frames of oil leakage region boundaries, and establishing an oil leakage diffusion trend prediction curve according to the oil leakage diffusion speed data; Obtaining the surface image of the base seal ring corresponding to the oil leakage spectrogram, and identifying the elastic deformation information of the base seal ring in the surface image by using an edge detection algorithm; Irradiate the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; Obtain the third reflectance characteristics of the leakage oil spectrogram under multispectral imaging. Based on the third reflectance characteristics, the ultraviolet fluorescence reaction characteristics, the second reflectance characteristics, the leakage oil diffusion trend prediction curve, and the elastic deformation information, perform gap detection training on the gap detection model.

8. A risk detection device for generating gaps in substation transformers, characterized in that, It includes: A feature extraction module for obtaining the first reflectance characteristics of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer under multispectral imaging, and obtaining the first reflectance change map of the base seal ring and the metal contact surface according to the first reflectance characteristics of different regions; A model construction module for constructing a gap detection model according to the spectral response differences of the first reflectance change map under different elastic deformation characteristics of the base seal ring; A region division module for dividing the regions with a change speed greater than a preset speed into gap generation regions according to the change speed of the reflectance characteristics of the first reflectance change map, obtaining the second reflectance characteristics of the gap generation regions under multispectral imaging, and pre-evaluating the gap generation risk according to the second reflectance characteristics to obtain a first risk detection result; An ultraviolet reaction module for obtaining a leakage oil sample dataset of the transformer, and irradiating the leakage oil sample images in the leakage oil sample dataset with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; A model training module for performing gap detection training on the gap detection model according to the third reflectance characteristics, the ultraviolet fluorescence reaction characteristics, and the second reflectance characteristics of the leakage oil sample images in the leakage oil sample dataset; A risk detection module for inputting the transformer image data collected in real time into the trained gap detection model during the actual operation of the substation to obtain a second risk detection result, and detecting and evaluating the gap generation risk of the target transformer according to the first risk detection result and the second risk detection result.

9. The risk detection device for the substation transformer mutual inductor gap generation according to claim 8, characterized in that, The feature extraction module is used for: Obtain the spectral images of different regions at the edge position of the base seal ring and the metal contact surface of the substation transformer under multispectral imaging, and extract the first spectral reflectance data from the spectral images; Preprocess the first spectral reflectance data using a Gaussian filter, and extract the first reflectance characteristics from the preprocessed first spectral reflectance data; Use the reflectance difference calculation method to calculate the difference between the first reflectance characteristics at adjacent positions in the image, and obtain a first reflectance change map reflecting the change degree of the reflectance within a unit distance.

10. The risk detection device for substation transformer gap generation according to claim 8, characterized in that, The model construction module is used for: Obtain a sample set of elastic deformation data of the base seal ring when there is a gap in the transformer; Extract several elastic deformation characteristics of the base seal ring when there is a gap in the transformer according to the elastic deformation data sample set; Determine the mapping relationship between the elastic deformation characteristics and the gap generation conditions according to each of the elastic deformation characteristics and the corresponding first reflectivity change map, and construct a gap detection model according to the mapping relationship.

11. The risk detection device for substation transformer mutual inductor gap generation according to claim 8, characterized in that, The region division module is used for: Extract the reflectivity jump data with the change speed of the reflectivity characteristics greater than the preset speed in the first reflectivity change map, and use the maximum inter-class variance method to segment the reflectivity jump data to obtain the gap candidate regions where the reflectivity characteristics change suddenly; Adjust the division boundaries of the gap candidate regions according to the region area, edge trend and spectral intensity of the gap candidate regions to obtain the gap generation regions and obtain the second reflectivity characteristics of the gap generation regions under multi-spectral imaging.

12. The risk detection device for the substation transformer mutual inductor gap generation according to claim 8, characterized in that, The device further includes a risk judgment module: Calculate the change rate of the second reflectivity characteristics within the gap generation region to obtain the second reflectivity change map within the gap generation region; Screen the second reflectivity characteristics of the gap generation region according to the second reflectivity change map. If there is a point where the change rate value of the second reflectivity characteristics within the gap generation region is greater than the preset reflectivity change threshold, it is determined that there is a risk of gap generation in the gap generation region corresponding to this point.

13. The risk detection device for the substation transformer mutual inductor gap generation as described in claim 12, wherein The reflectivity change threshold is a multi-level reflectivity change threshold; The risk judgment module is further used for: The step of determining that there is a risk of gap generation in the gap generation region corresponding to a point if there is a point where the change rate value of the second reflectivity characteristics within the gap generation region is greater than the preset reflectivity change threshold includes: If there is a point where the change rate value of the second reflectivity characteristics within the gap generation region is greater than the preset reflectivity change threshold, classify the risk level of gap generation in the gap generation region according to the change rate value and the multi-level reflectivity change threshold.

14. The risk detection device for the substation transformer mutual inductor gap generation as described in claim 8, wherein, The model training module is used for: Obtain the mutual inductor oil leakage sample data set, and obtain the oil leakage spectrograms of each oil leakage sample image in the oil leakage sample data set under multi-spectral imaging; Extract the oil leakage distribution characteristics in the oil leakage spectrogram, and determine the oil leakage region boundary in the oil leakage spectrogram according to the oil leakage distribution characteristics; Calculate the oil leakage diffusion speed data by using the optical flow method based on at least two consecutive frames of oil leakage region boundaries, and establish an oil leakage diffusion trend prediction curve according to the oil leakage diffusion speed data; Obtain the surface image of the base sealing ring corresponding to the oil leakage spectrogram, and use an edge detection algorithm to identify the elastic deformation information of the base sealing ring in the surface image; Irradiate the oil leakage sample images in the oil leakage sample data set with an ultraviolet simulation light source to obtain ultraviolet fluorescence reaction characteristics; Obtain the third reflectivity characteristics of the oil leakage spectrogram under multi-spectral imaging, and perform gap detection training on the gap detection model according to the third reflectivity characteristics, the ultraviolet fluorescence reaction characteristics, the second reflectivity characteristics, the oil leakage diffusion trend prediction curve and the elastic deformation information.

15. A computer device, characterized in that, It includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the risk detection method for substation transformer mutual inductor gap generation as described in any one of claims 1 to 7.

16. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program. When the device where the computer-readable storage medium is located executes the computer program, it implements the risk detection method for substation transformer mutual inductor gap generation as described in any one of claims 1 to 7.

Images