Power transmission line infrared thermal hidden danger and fault identification method and system and storage medium

Abstract

The embodiment of the invention discloses a power transmission line infrared thermal hidden danger and fault identification method and system and a storage medium, and the method comprises the steps: synchronously collecting an infrared thermal image, a visible light image, environment condition data and operation parameters of a typical part in a power transmission line, and carrying out the alignment of the infrared thermal image and the visible light image; obtaining thermal radiation data of typical parts in the aligned infrared thermal image through a target detection model; fusing the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector; constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identifying an infrared thermal hidden danger area of a typical part in the power transmission line through the thermal radiation intensity and temperature model; and fusing the thermal radiation data of each typical component with the operation parameters to generate a second comprehensive feature vector, thereby identifying potential fault points of the typical components in the power transmission line, discovering equipment fault signs in advance, and improving the operation safety and reliability of the power system.

Description

Technical Field

[0001] The present invention relates to the technical field of power transmission line monitoring, and in particular to a method, system and storage medium for identifying infrared thermal hazards and faults of a power transmission line. Background Art

[0002] In the power system, typical components of transmission lines such as insulators and wire clamps are exposed to various external environments for a long time and are easily affected by factors such as temperature, humidity, and wind speed, leading to problems such as overheating and aging. These problems will not only affect the normal operation of the power system, but may also cause serious safety accidents.

[0003] At present, the commonly used transmission line monitoring methods mainly include temperature monitoring based on single-modal data and rule-based fault diagnosis. Using only infrared images or a single type of operating parameters cannot fully reflect the operating status of components and is prone to missing key information. For example, infrared images can only provide temperature distribution, but cannot capture the structural characteristics of components, which are equally important for fault diagnosis. Fault diagnosis methods based on preset thresholds rely too much on manually set thresholds and lack adaptability and flexibility. Under different environmental conditions, thresholds need to be adjusted frequently, otherwise it is easy to cause false alarms or missed alarms, and it is difficult to cope with complex and changing operating environments. In addition, there are some machine learning-based methods that train models with historical data to predict the health status of components. Although these methods can capture some complex patterns, when the amount of data processed is large, the generalization ability of the model is poor, and it is easy to have overfitting or underfitting problems. In addition, these methods usually require a large amount of labeled data, and in practical applications, the cost of obtaining labeled data is high. Summary of the invention

[0004] Based on this, it is necessary to propose a method, system and storage medium for identifying infrared thermal hazards and faults of transmission lines in response to the above problems.

[0005] A method for identifying infrared thermal hazards and faults of a power transmission line, the method comprising:

[0006] Infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line are collected synchronously, and the infrared thermal images and visible light images are aligned.

[0007] The thermal radiation data of typical components in the aligned infrared thermal images are acquired through the target detection model.

[0008] The thermal radiation data of each typical component is fused with the environmental condition data to generate a first comprehensive feature vector.

[0009] A thermal radiation intensity and temperature model is constructed according to the first comprehensive feature vector, and infrared heat hazard areas of typical components in the transmission line are identified through the thermal radiation intensity and temperature model.

[0010] The thermal radiation data of each typical component is fused with the operating parameters to generate a second comprehensive feature vector.

[0011] Potential fault points of typical components in the transmission line are identified according to the second comprehensive feature vector.

[0012] The synchronous acquisition of infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line, and the alignment of the infrared thermal images and visible light images in time and space specifically includes:

[0013] Synchronously collect infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line.

[0014] The infrared thermal image and the visible light image are aligned in time through a time stamp.

[0015] Matching points with the same position information are marked on the infrared thermal image and the visible light image respectively.

[0016] A geometric transformation matrix between the infrared thermal image and the visible light image is determined according to the matching points.

[0017] The infrared thermal image is transformed by the geometric transformation matrix to be spatially aligned with the visible light image.

[0018] The step of acquiring the thermal radiation data of the corresponding components in the aligned infrared thermal images through the target detection model specifically includes:

[0019] The bounding box of each typical component in the aligned visible light image is extracted by the object detection model.

[0020] Based on the bounding box of each typical component, thermal radiation data of the corresponding component is extracted from the aligned infrared thermal images.

[0021] The step of constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector and identifying infrared heat hazard areas of typical components in the power transmission line through the thermal radiation intensity and temperature model specifically includes:

[0022] A thermal radiation intensity and temperature model is constructed based on the first comprehensive feature vector.

[0023] The current thermal radiation data is converted into a current temperature value through the thermal radiation intensity and temperature model.

[0024] According to the comparison between the current temperature value and the temperature threshold, infrared heat hazard areas of typical components in the power transmission line are identified.

[0025] The step of identifying infrared heat hazard areas of typical components in the power transmission line according to the comparison between the current temperature value and the temperature threshold value specifically includes:

[0026] It is determined whether the current temperature value is greater than a temperature threshold.

[0027] If the current temperature value is greater than the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as an infrared heat hazard area.

[0028] If the current temperature value is less than or equal to the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as a normal area.

[0029] The step of identifying potential fault points of typical components in the transmission line according to the second comprehensive feature vector specifically includes:

[0030] A fault prediction model is constructed with the second comprehensive feature vector as input and potential fault points of typical components in the transmission line as prediction values, and a priori distribution is introduced on model parameters of the fault prediction model using a Bayesian neural network.

[0031] The current data points in the second comprehensive feature vector are input into the fault prediction model to identify potential fault points of typical components in the transmission line and uncertain values ​​of current prediction values.

[0032] If the uncertainty value of the current prediction value is determined to be greater than a preset uncertainty threshold, the confidence of the fault prediction model is reduced.

[0033] If the uncertainty value of the current prediction value is determined to be less than or equal to a preset uncertainty threshold, the confidence of the fault prediction model is improved.

[0034] The step of identifying potential fault points of typical components in the transmission line according to the second comprehensive feature vector specifically includes:

[0035] Clustering is performed on each data point in the second comprehensive feature vector to determine an optimal number of clusters.

[0036] Each data point is divided into different clusters according to the optimal number of clusters.

[0037] The characteristic value of each data point in each cluster is extracted, and the abnormal score index of the data point is determined according to the characteristic value.

[0038] Potential fault points of typical components in the transmission line are identified based on the comparison between the data point anomaly score index and the anomaly index threshold.

[0039] The step of clustering each data point in the second comprehensive feature vector to determine the optimal number of clusters specifically includes:

[0040] A preset number of clusters is set to cluster each data point in the second comprehensive feature vector.

[0041] For each of the preset initial clusters, a silhouette coefficient is determined according to the average distance of each data point to other data points in the same cluster and the average distance to all data points in the nearest cluster.

[0042] By comparing the silhouette coefficients under different preset cluster numbers, the cluster number with the largest silhouette coefficient is selected as the optimal cluster number.

[0043] The step of clustering each data point in the second comprehensive feature vector to determine the optimal number of clusters specifically includes:

[0044] A preset number of clusters is set to cluster each data point in the second comprehensive feature vector.

[0045] By comparing the clustering errors under different preset cluster numbers, the cluster number with the smallest clustering error is selected as the optimal cluster number.

[0046] The step of extracting the characteristic value of each data point in each cluster and determining the data point abnormality score index according to the characteristic value specifically includes:

[0047] The characteristic value of each data point in each cluster is extracted, and an abnormal score index model is determined according to the characteristic value. The abnormal score index model is:

[0048]

[0049] Among them, x i is the eigenvalue of the i-th data point, μ cluster is the mean of the cluster, σ cluster is the standard deviation of the cluster, α is the weight parameter, f(x i ) is a nonlinear function, f(x i )About x i The partial derivative of .

[0050] The feature value of the current data point is input into the anomaly score index model to determine the anomaly score index of the current data point.

[0051] The step of identifying potential fault points of typical components in the power transmission line according to the comparison between the abnormal score index of the data point and the abnormal index threshold specifically includes:

[0052] Determine whether the anomaly score indicator of the current data point is greater than the anomaly threshold.

[0053] If the anomaly score indicator of the current data point is greater than the anomaly index threshold, the current data point is marked as a potential fault point.

[0054] If the anomaly score indicator of the current data point is less than or equal to the anomaly index threshold, the current data point is marked as a normal fault point.

[0055] A transmission line infrared thermal hazard and fault identification system, the system comprising:

[0056] The infrared thermal image and visible light image alignment module is used to synchronously collect infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line, and align the infrared thermal images and visible light images.

[0057] The thermal radiation data acquisition module is used to acquire the thermal radiation data of typical components in the aligned infrared thermal image through a target detection model.

[0058] The first comprehensive feature vector generating module is used to fuse the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector.

[0059] The infrared heat potential hazard area identification module of typical components in the transmission line is used to construct a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identify the infrared heat potential hazard area of ​​typical components in the transmission line through the thermal radiation intensity and temperature model.

[0060] The second comprehensive feature vector generating module is used to fuse the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector.

[0061] A potential fault point identification module for typical components in a transmission line is used to identify potential fault points of typical components in a transmission line according to the second comprehensive feature vector.

[0062] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the processor executes the steps of the method described above.

[0063] The embodiments of the present invention have the following beneficial effects:

[0064] The present invention aligns the infrared thermal image with the visible light image so that the thermal radiation information in the infrared image corresponds to the actual scene in the visible light image, which not only retains the thermal radiation information, but also contains the structural features of the typical components in the transmission line, providing rich information for subsequent infrared thermal hazards and fault identification. The target detection model identifies and extracts the thermal radiation data of the typical components from the aligned infrared thermal image, fuses the thermal radiation data of each typical component with the environmental condition data, generates a first comprehensive feature vector, and constructs a thermal radiation intensity and temperature model based on the first comprehensive feature vector, thereby realizing efficient detection and early warning of heating hazards of typical components of the transmission line, and being able to timely discover abnormal thermal radiation areas, providing important guarantees for the maintenance and safe operation of the equipment. In addition, the thermal radiation data of each typical component is fused with the operating parameters to generate a second comprehensive feature vector, and according to the second comprehensive feature vector, the potential fault points of the typical components in the transmission line are identified, and the fault signs of the equipment are discovered in advance, which greatly improves the operational safety and reliability of the power system and reduces the economic losses and safety risks caused by potential faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0066] in:

[0067] Figure 1 A schematic diagram of a flow chart of an embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention;

[0068] Figure 2 A schematic flow chart of another embodiment of a method for identifying infrared heat hazards and faults of power transmission lines provided by the present invention;

[0069] Figure 3 A schematic flow chart of another embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention;

[0070] Figure 4 A schematic flow chart of another embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention;

[0071] Figure 5 A schematic diagram of the structure of an embodiment of a power transmission line infrared thermal hazard and fault identification system provided by the present invention;

[0072] Figure 6A schematic structural diagram of an embodiment of a storage medium provided by the present invention. DETAILED DESCRIPTION

[0073] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0074] like Figure 1 As shown, Figure 1 A schematic diagram of a flow chart of an embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention. A method for identifying infrared heat hazards and faults of a power transmission line, the method comprising:

[0075] S101: synchronously collect infrared thermal images, visible light images, environmental condition data, and operating parameters of typical components in the transmission line, and align the infrared thermal images and visible light images.

[0076] In a specific implementation scenario, a dual-light imaging system consisting of a high-resolution infrared thermal imager and a visible light camera is used to synchronously collect infrared thermal images and visible light images of target electrical components. The high-resolution infrared thermal imager and the visible light camera are precisely aligned through a time synchronization module and a spatial alignment module to ensure that the infrared thermal image and the visible light image collected at the same time point and the same spatial position are consistent. Specifically, the built-in timestamp synchronization module is used to ensure that the high-resolution infrared thermal imager and the visible light camera are triggered to shoot at the same time point to avoid data inconsistency caused by time deviation. The two images are spatially aligned through a calibration plate, which contains marking points with known positions. Matching points with the same position information are marked on the infrared thermal image and the visible light image respectively; the geometric transformation matrix between the infrared thermal image and the visible light image is determined based on the matching points; the infrared thermal image is transformed through the geometric transformation matrix, and is spatially aligned with the visible light image to achieve pixel-level alignment.

[0077] S102: Acquire thermal radiation data of typical components in the aligned infrared thermal image through the target detection model.

[0078] In a specific implementation scenario, advanced target detection models such as YOLOv5, Faster R-CNN, etc. are selected. These target detection models have high accuracy and real-time performance and are suitable for target detection tasks in complex backgrounds. A large number of visible light images containing typical components such as insulators and wire clamps are annotated to generate a training data set with bounding box labels. The target detection model is trained using the training data set to optimize the detection accuracy and robustness of the model. The aligned visible light images collected in real time are used for target detection through the trained target detection model to generate a bounding box for each component. Redundant detection results are removed through post-processing techniques such as non-maximum suppression to ensure that each component has only one bounding box. Furthermore, the generated bounding box is used to extract the thermal radiation data of typical components from the aligned infrared thermal images. The extracted thermal radiation data includes the temperature value of each pixel, forming a two-dimensional temperature matrix.

[0079] S103: Fusing the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector.

[0080] In a specific implementation scenario, the thermal radiation data and environmental condition data of each typical component extracted are fused to generate a first comprehensive feature vector. At the same time, in the fusion process, data preprocessing techniques such as normalization and standardization can be used to eliminate dimensional effects and improve the accuracy of subsequent analysis.

[0081] S104: constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identifying infrared heat hazard areas of typical components in the transmission line through the thermal radiation intensity and temperature model.

[0082] In a specific implementation scenario, a thermal radiation intensity and temperature model is constructed based on the first comprehensive feature vector, and the current thermal radiation data is converted into a current temperature value through the thermal radiation intensity and temperature model, and the infrared heat hazard area of ​​a typical component in the transmission line is identified based on the comparison between the current temperature value and the temperature threshold. Specifically, by setting a temperature threshold, the pixel points whose current temperature value exceeds the temperature threshold can be marked.

[0083] The present invention can also combine the historical temperature values ​​and environmental condition data of typical components, and use machine learning models (such as decision trees, random forests, etc.) to evaluate the risk level of infrared heat hazard areas. Among them, the risk level can be divided into three levels: low, medium, and high, which are used to guide the decision-making of operation and maintenance personnel.

[0084] S105: Fusion the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector.

[0085] In a specific implementation scenario, the infrared thermal image analysis result is fused with the operating parameters from the data acquisition and monitoring control system, and a second comprehensive feature vector is generated using a data fusion algorithm, such as a Bayesian network, a support vector machine, etc.

[0086] S106: Identify potential fault points of typical components in the transmission line according to the second comprehensive feature vector.

[0087] In a specific implementation scenario, a fault prediction model is constructed with the second comprehensive feature vector as input and the potential fault points of typical components in the transmission line as prediction values, and a Bayesian neural network is used to introduce prior distribution on the model parameters of the fault prediction model; the current data points in the second comprehensive feature vector are input into the fault prediction model to identify the potential fault points of typical components in the transmission line and the uncertainty values ​​of the current prediction values; if it is determined that the uncertainty value of the current prediction value is greater than a preset uncertainty threshold, the confidence of the fault prediction model is reduced; if it is determined that the uncertainty value of the current prediction value is less than or equal to the preset uncertainty threshold, the confidence of the fault prediction model is increased.

[0088] In another specific implementation scenario, each data point in the second comprehensive feature vector is clustered to determine the optimal number of clusters; each data point is divided into different clusters according to the optimal number of clusters; the eigenvalue of each data point in each cluster is extracted, and the data point anomaly score index is determined based on the eigenvalue; and the potential fault points of typical components in the transmission line are identified based on the comparison between the data point anomaly score index and the anomaly index threshold.

[0089] From the above description, it can be seen that the present invention aligns the infrared thermal image with the visible light image so that the thermal radiation information in the infrared image corresponds to the actual scene in the visible light image, which not only retains the thermal radiation information, but also contains the structural characteristics of the typical components in the transmission line, providing rich information for subsequent infrared thermal hazards and fault identification. The target detection model identifies and extracts the thermal radiation data of the typical components from the aligned infrared thermal image, fuses the thermal radiation data of each typical component with the environmental condition data, generates a first comprehensive feature vector, and constructs a thermal radiation intensity and temperature model based on the first comprehensive feature vector, realizing efficient detection and early warning of heating hazards of typical components of the transmission line, and can timely discover abnormal thermal radiation areas, providing important guarantees for the maintenance and safe operation of the equipment. In addition, the thermal radiation data of each typical component is fused with the operating parameters to generate a second comprehensive feature vector. According to the second comprehensive feature vector, the potential fault points of the typical components in the transmission line are identified, and the fault signs of the equipment are discovered in advance, which greatly improves the operational safety and reliability of the power system and reduces the economic losses and safety risks caused by potential faults.

[0090] like Figure 2 As shown, Figure 2A schematic diagram of another embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention. A method for identifying infrared heat hazards and faults of a power transmission line, the method comprising:

[0091] S201: Synchronously collect infrared thermal images, visible light images, environmental condition data, and operating parameters of typical components in the transmission line.

[0092] In a specific implementation scenario, a dual-light imaging system consisting of a high-resolution infrared thermal imager and a visible light camera is used to synchronously capture infrared thermal images and visible light images of target electrical components.

[0093] S202: Temporally aligning the infrared thermal image and the visible light image through a timestamp.

[0094] S203: Mark matching points with the same position information on the infrared thermal image and the visible light image respectively.

[0095] S204: Determine a geometric transformation matrix between the infrared thermal image and the visible light image according to the matching points.

[0096] S205: Transform the infrared thermal image through a geometric transformation matrix to align it with the visible light image in space.

[0097] In a specific implementation scenario, the high-resolution infrared thermal imager and the visible light camera are precisely aligned through the time synchronization module and the spatial alignment module to ensure that the infrared thermal image and the visible light image collected at the same time point and the same spatial position are consistent. Specifically, the built-in timestamp synchronization module is used to ensure that the high-resolution infrared thermal imager and the visible light camera are triggered to shoot at the same time point to avoid data inconsistency caused by time deviation. The two images are spatially aligned through a calibration plate, which contains marking points with known positions. Matching points with the same position information are marked on the infrared thermal image and the visible light image respectively; the geometric transformation matrix between the infrared thermal image and the visible light image is determined based on the matching points; the infrared thermal image is transformed through the geometric transformation matrix, and is spatially aligned with the visible light image to achieve pixel-level alignment.

[0098] S206: Extracting the bounding box of each typical component in the aligned visible light image through the object detection model.

[0099] S207: Based on the bounding box of each typical component, extract thermal radiation data of the corresponding component from the aligned infrared thermal image.

[0100] S208: Fusing the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector.

[0101] It should be noted that steps S206-S208 are Figure 1 This has been discussed in detail in the implementation scenario shown and will not be repeated here.

[0102] S209: Constructing a thermal radiation intensity and temperature model according to the first comprehensive eigenvector.

[0103] In a specific implementation scenario, based on the first comprehensive feature vector, a nonlinear regression model (such as support vector regression SVR, random forest, etc.) is used to construct a thermal radiation intensity and temperature curve, that is, a thermal radiation intensity and temperature model. In addition, during the model training process, cross-validation technology is used to optimize hyperparameters to ensure the generalization ability and prediction accuracy of the model.

[0104] S210: Convert the current thermal radiation data into a current temperature value through a thermal radiation intensity and temperature model.

[0105] S211: Determine whether the current temperature value is greater than the temperature threshold.

[0106] S2111: If the current temperature value is greater than the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as an infrared heat hazard area.

[0107] S2112: If the current temperature value is less than or equal to the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as a normal area.

[0108] In one implementation scenario, the current thermal radiation data is converted into the current temperature value using the thermal radiation intensity and temperature model. By setting the temperature threshold, the pixels whose temperature exceeds the threshold are automatically marked. Specifically, it is determined whether the current temperature value is greater than the temperature threshold. If the current temperature value is greater than the temperature threshold, the pixel corresponding to the current thermal radiation data is marked as an infrared heat hazard area. If the current temperature value is less than or equal to the temperature threshold, the pixel corresponding to the current thermal radiation data is marked as a normal area.

[0109] S212: Fusion the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector.

[0110] S213: Identify potential fault points of typical components in the transmission line according to the second comprehensive feature vector.

[0111] It should be noted that steps S212-S213 are Figure 1 This has been discussed in detail in the implementation scenario shown and will not be repeated here.

[0112] like Figure 3 As shown, Figure 3A schematic flow chart of another embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention. A method for identifying infrared heat hazards and faults of a power transmission line, the method comprising:

[0113] S301: Taking the second comprehensive feature vector as input and the potential fault points of typical components in the transmission line as prediction values, a fault prediction model is constructed, and a Bayesian neural network is used to introduce a priori distribution on model parameters of the fault prediction model.

[0114] In one implementation scenario, the second comprehensive feature vector is used as input, and the potential fault points of typical components in the transmission line are used as prediction values ​​to construct a fault prediction model, and a Bayesian deep learning method, such as a Bayesian neural network, is introduced to perform uncertainty quantitative analysis on the prediction results of the fault prediction model. Bayesian deep learning can estimate the posterior distribution of the model parameters by introducing a priori distribution on the model parameters, thereby providing an uncertainty value of the prediction value. In the model prediction process, not only the prediction value is output, but also the uncertainty measure of the prediction value, the standard deviation of the prediction value or the confidence interval.

[0115] S302: The current data point in the second comprehensive feature vector is input into the fault prediction model to identify potential fault points of typical components in the transmission line and uncertain values ​​of current prediction values.

[0116] S303: If it is determined that the uncertainty value of the current prediction value is greater than a preset uncertainty threshold, the confidence of the fault prediction model is reduced.

[0117] S304: If it is determined that the uncertainty value of the current prediction value is less than or equal to a preset uncertainty threshold, the confidence of the fault prediction model is increased.

[0118] In one implementation scenario, the current data point in the second comprehensive feature vector is input into the fault prediction model to identify the potential fault points of typical components in the transmission line and the uncertainty value of the current prediction value. The confidence of the fault prediction model can be evaluated by the uncertainty value of the current prediction value. The confidence level of fault identification is dynamically adjusted according to the comparison between the uncertainty value of the prediction value and the preset uncertainty threshold. Specifically, if it is determined that the uncertainty value of the current prediction value is greater than the preset uncertainty threshold, the confidence of the fault prediction model is reduced and the need for further verification is increased. On the contrary, if it is determined that the uncertainty value of the current prediction value is less than or equal to the preset uncertainty threshold, the confidence of the fault prediction model is increased and false alarms are reduced.

[0119] From the above description, it can be seen that the present invention introduces the Bayesian deep learning method to perform uncertainty quantification analysis and dynamically adjust the confidence level of fault prediction.

[0120] like Figure 4 As shown, Figure 4A schematic flow chart of another embodiment of a method for identifying infrared heat hazards and faults of a power transmission line provided by the present invention. A method for identifying infrared heat hazards and faults of a power transmission line, the method comprising:

[0121] S401: Clustering each data point in the second comprehensive feature vector to determine the optimal number of clusters.

[0122] In one implementation scenario, an unsupervised learning method suitable for high-dimensional data is selected for cluster analysis, such as Gaussian mixture model, spectral clustering or deep embedding clustering. These methods can capture the complex distribution structure of data and are suitable for cluster analysis of multimodal data. The selected unsupervised learning method is used to perform cluster analysis on the comprehensive feature vector. Specifically, a preset number of clusters is set, and each data point in the second comprehensive feature vector is clustered; for each preset initial cluster, the silhouette coefficient is determined based on the average distance of each data point to other data points in the same cluster and the average distance to all data points in the nearest cluster. The silhouette coefficient is:

[0123]

[0124] Where a(i) is the average distance between data point i and other points in its cluster, b(i) is the average distance between data point i and all points in its nearest clusters, β is the weight parameter of the numerator, which is used to adjust the influence of the overall numerator. Its value range is usually [0,1], and its default value is 1. p is the power parameter of the numerator, which is used to control the nonlinear transformation of b(i)-a(i). Its value range is usually [1,3], and its default value is 1. γ is the weight parameter of the a(i) part in the denominator, which is used to adjust the influence of a(i). Its value range is usually [0,1], and its default value is 1. q is the power parameter of the a(i) part in the denominator, which is used to control the nonlinear transformation of a(i). Its value range is usually [1,3], and its default value is 1. δ is the weight parameter of the b(i) part in the denominator, which is used to adjust the influence of b(i). Its value range is usually [0,1], and its default value is 1. r is the power parameter of the b(i) part in the denominator, which is used to control the nonlinear transformation of b(i). Its value range is usually [1,3], and its default value is 1. λ is the sigmoid function parameter of the a(i) part in the denominator, which is used to adjust the nonlinear transformation of a(i). Its value range is usually [0,10], and its default value is 1. μ is the offset parameter of the sigmoid function of the a(i) part in the denominator, which is used to adjust the nonlinear transformation of a(i). Its value range is usually [0,10], and its default value is 1. η is the sigmoid function parameter of the b(i) part in the denominator, which is used to adjust the nonlinear transformation of b(i). Its value range is usually [0,10], and its default value is 1. ν is the offset parameter of the sigmoid function of the b(i) part in the denominator, which is usually [0,10], and its default value is 1.

[0125] Furthermore, by comparing the silhouette coefficients under different preset cluster numbers, the cluster number with the largest silhouette coefficient is selected as the optimal cluster number.

[0126] In another implementation scenario, a preset number of clusters is set, and each data point in the second comprehensive feature vector is clustered; by comparing the clustering errors under different preset numbers of clusters, the number of clusters with the smallest clustering error is selected as the optimal number of clusters.

[0127] S402: Divide each data point into different clusters according to the optimal number of clusters.

[0128] S403: Extracting the characteristic value of each data point in each cluster, and determining the data point abnormality score index according to the characteristic value.

[0129] In an implementation scenario, the characteristic value of each data point in each cluster is extracted, and the anomaly score indicator model is determined according to the characteristic value. The anomaly score indicator model is:

[0130]

[0131] Among them, x i is the eigenvalue of the i-th data point, μ cluster is the mean of the cluster, σ cluster is the standard deviation of the cluster, α is the weight parameter used to adjust the influence of the derivative term, f(x i ) is a nonlinear function used to capture the local behavior of data points in the feature space. f(x i )About x i The partial derivative of is used to measure the rate of change of data points in the feature space.

[0132] Furthermore, the feature value of the current data point is input into the anomaly score index model to determine the anomaly score index of the current data point.

[0133] S404: Determine whether the anomaly score index of the current data point is greater than the anomaly index threshold.

[0134] S4041: If the anomaly score index of the current data point is greater than the anomaly index threshold, the current data point is marked as a potential fault point.

[0135] S4042: If the anomaly score index of the current data point is less than or equal to the anomaly index threshold, the current data point is marked as a normal fault point.

[0136] In one implementation scenario, it is determined whether the anomaly score index of the current data point is greater than the anomaly index threshold. If the anomaly score index of the current data point is greater than the anomaly index threshold, the current data point is marked as a potential fault point, and if the anomaly score index of the current data point is less than or equal to the anomaly index threshold, the current data point is marked as a normal fault point.

[0137] From the above description, it can be seen that the present invention extracts the eigenvalues ​​of each data point in each cluster and determines the anomaly score index model according to the eigenvalues. It not only considers the standardized distance of the data point relative to the cluster to which it belongs, but also captures the local behavior and change rate of the data point in the feature space through nonlinear functions and their derivatives, which significantly improves the accuracy and robustness of anomaly detection, thereby effectively enhancing the early identification and warning capabilities of potential faults in the power system.

[0138] like Figure 5 As shown, Figure 5 The present invention provides a schematic diagram of the structure of a transmission line infrared heat hazard and fault identification system according to an embodiment of the present invention. A transmission line infrared heat hazard and fault identification system 10, the system comprising:

[0139] The infrared thermal image and visible light image alignment module 11 is used to synchronously collect infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line, and align the infrared thermal images and visible light images.

[0140] The thermal radiation data acquisition module 12 is used to acquire the thermal radiation data of typical components in the aligned infrared thermal image through a target detection model.

[0141] The first comprehensive feature vector generating module 13 is used to fuse the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector.

[0142] The infrared heat potential hazard area identification module 14 of typical components in the transmission line is used to construct a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identify the infrared heat potential hazard area of ​​typical components in the transmission line through the thermal radiation intensity and temperature model.

[0143] The second comprehensive feature vector generating module 15 is used to fuse the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector.

[0144] The potential fault point identification module 16 of the typical components in the power transmission line is used to identify the potential fault point of the typical components in the power transmission line according to the second comprehensive feature vector.

[0145] In one implementation scenario, in the infrared thermal image and visible light image alignment module 11, the infrared thermal image, visible light image, environmental condition data and operating parameters of typical components in the transmission line are collected synchronously; the infrared thermal image and the visible light image are aligned in time through timestamps; matching points with the same position information are marked on the infrared thermal image and the visible light image respectively; the geometric transformation matrix between the infrared thermal image and the visible light image is determined according to the matching points; the infrared thermal image is transformed by the geometric transformation matrix and aligned with the visible light image in space. In the thermal radiation data acquisition module 12, the bounding box of each typical component in the aligned visible light image is extracted through the target detection model; based on the bounding box of each typical component, the thermal radiation data of the corresponding component is extracted from the aligned infrared thermal image. In the first comprehensive feature vector generation module 13, the thermal radiation data of each typical component is fused with the environmental condition data to generate a first comprehensive feature vector. In the infrared heat hazard area identification module 14 of typical components in the transmission line, a thermal radiation intensity and temperature model is constructed according to the first comprehensive feature vector; the current thermal radiation data is converted into the current temperature value through the thermal radiation intensity and temperature model; and the infrared heat hazard area of ​​the typical components in the transmission line is identified according to the comparison between the current temperature value and the temperature threshold. In the second comprehensive feature vector generation module 15, the thermal radiation data of each typical component is fused with the operating parameters to generate a second comprehensive feature vector. In the potential fault point identification module 16 of typical components in the transmission line, the potential fault points of the typical components in the transmission line are identified according to the second comprehensive feature vector.

[0146] like Figure 6 As shown, Figure 6 The storage medium 20 stores at least one computer program 21, which is executed by a processor to implement the following. Figure 1 , Figure 2 , Figure 3 and Figure 4 In one embodiment, the storage medium 20 may be a storage chip, a hard disk, a mobile hard disk, a USB flash drive, an optical disk, or other readable and writable storage tools, or a server, etc.

[0147] The above describes specific embodiments of the present specification, and other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily have to be performed in the specific order or sequential order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0148] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, equipment, and non-volatile computer-readable storage medium embodiments, since they are basically similar to the method embodiments, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiments.

[0149] The apparatus, device, non-volatile computer-readable storage medium and method provided in the embodiments of this specification correspond to each other, and therefore, the apparatus, device, and non-volatile computer storage medium also have similar beneficial technical effects as the corresponding methods. Since the beneficial technical effects of the methods have been described in detail above, the beneficial technical effects of the corresponding apparatus, device, and non-volatile computer storage medium will not be repeated here.

[0150] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0151] For the convenience of description, the above device is described by being divided into various units according to their functions and described separately. Of course, when implementing this specification, the functions of each unit can be implemented in one or more software and / or hardware. It should be understood by those skilled in the art that this specification embodiment can be provided as a method, system, or computer program product. Therefore, this specification embodiment can adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification embodiment can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0153] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0154] The above disclosure is only the preferred embodiment of the present invention, which certainly cannot be used to limit the scope of the present invention. Therefore, equivalent changes made according to the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for identifying infrared thermal hazards and faults in power transmission lines, characterized in that: The method comprises: Synchronously collect infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line, and align the infrared thermal images and visible light images; Acquiring thermal radiation data of typical components in the aligned infrared thermal images through a target detection model; fusing the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector; Constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identifying infrared heat hazard areas of typical components in the transmission line through the thermal radiation intensity and temperature model; fusing the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector; Potential fault points of typical components in the transmission line are identified according to the second comprehensive feature vector.

2. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 1, characterized in that: The synchronous acquisition of infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line, and alignment of the infrared thermal images and visible light images in time and space specifically includes: Simultaneously collect infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the transmission line; Temporally aligning the infrared thermal image and the visible light image by using a timestamp; Marking matching points with the same position information on the infrared thermal image and the visible light image respectively; Determine a geometric transformation matrix between the infrared thermal image and the visible light image according to the matching points; The infrared thermal image is transformed by the geometric transformation matrix to be spatially aligned with the visible light image.

3. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 2, characterized in that: The step of acquiring the thermal radiation data of the corresponding components in the aligned infrared thermal images through the target detection model specifically includes: Extracting a bounding box of each typical component in the aligned visible light image through a target detection model; Based on the bounding box of each typical component, thermal radiation data of the corresponding component is extracted from the aligned infrared thermal images.

4. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 3, characterized in that: The step of constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector and identifying infrared heat hazard areas of typical components in the power transmission line through the thermal radiation intensity and temperature model specifically includes: Constructing a thermal radiation intensity and temperature model according to the first comprehensive feature vector; Converting the current thermal radiation data into the current temperature value through the thermal radiation intensity and temperature model; According to the comparison between the current temperature value and the temperature threshold, infrared heat hazard areas of typical components in the power transmission line are identified.

5. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 4, characterized in that: The step of identifying infrared heat hazard areas of typical components in the power transmission line according to the comparison between the current temperature value and the temperature threshold value specifically includes: Determine whether the current temperature value is greater than a temperature threshold; If the current temperature value is greater than the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as an infrared heat hazard area; If the current temperature value is less than or equal to the temperature threshold, the pixel point corresponding to the current thermal radiation data is marked as a normal area.

6. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 2, characterized in that: The step of identifying potential fault points of typical components in the transmission line according to the second comprehensive feature vector specifically includes: Taking the second comprehensive feature vector as input and the potential fault points of typical components in the transmission line as prediction values, a fault prediction model is constructed, and a prior distribution is introduced on the model parameters of the fault prediction model using a Bayesian neural network; The current data point in the second comprehensive feature vector is input into the fault prediction model to identify potential fault points of typical components in the transmission line and uncertain values ​​of the current prediction value; If it is determined that the uncertainty value of the current prediction value is greater than a preset uncertainty threshold, the confidence of the fault prediction model is reduced; If the uncertainty value of the current prediction value is determined to be less than or equal to a preset uncertainty threshold, the confidence of the fault prediction model is improved.

7. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 2, characterized in that: The step of identifying potential fault points of typical components in the transmission line according to the second comprehensive feature vector specifically includes: Clustering each data point in the second comprehensive feature vector to determine an optimal number of clusters; Divide each data point into different clusters according to the optimal number of clusters; Extracting the characteristic value of each data point in each cluster, and determining the data point abnormal score index according to the characteristic value; Potential fault points of typical components in the transmission line are identified based on the comparison between the data point anomaly score index and the anomaly index threshold.

8. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 7, characterized in that: The clustering of each data point in the second comprehensive feature vector to determine the optimal number of clusters specifically includes: Setting a preset number of clusters to cluster each data point in the second comprehensive feature vector; For each of the preset initial clusters, determining a silhouette coefficient according to an average distance of each data point to other data points in the same cluster and an average distance to all data points in the nearest cluster; By comparing the silhouette coefficients under different preset cluster numbers, the cluster number with the largest silhouette coefficient is selected as the optimal cluster number.

9. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 7, characterized in that: The clustering of each data point in the second comprehensive feature vector to determine the optimal number of clusters specifically includes: Setting a preset number of clusters to cluster each data point in the second comprehensive feature vector; By comparing the clustering errors under different preset cluster numbers, the cluster number with the smallest clustering error is selected as the optimal cluster number.

10. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 7, characterized in that: The extracting the characteristic value of each data point in each cluster and determining the data point abnormal score index according to the characteristic value specifically includes: The characteristic value of each data point in each cluster is extracted, and an abnormal score index model is determined according to the characteristic value. The abnormal score index model is: Among them, x i is the eigenvalue of the i-th data point, μ cluster is the mean of the cluster, σ cluster is the standard deviation of the cluster, α is the weight parameter, f(x i ) is a nonlinear function, f(x i )About x i The partial derivative of The feature value of the current data point is input into the anomaly score index model to determine the anomaly score index of the current data point.

11. The method for identifying infrared heat hazards and faults of power transmission lines according to claim 7, characterized in that: The step of identifying potential fault points of typical components in the power transmission line according to the comparison between the abnormal score index of the data point and the abnormal index threshold specifically includes: Determine whether the anomaly score indicator of the current data point is greater than the anomaly threshold; If the anomaly score index of the current data point is greater than the anomaly index threshold, the current data point is marked as a potential fault point; If the anomaly score indicator of the current data point is less than or equal to the anomaly index threshold, the current data point is marked as a normal fault point.

12. A power transmission line infrared thermal hazard and fault identification system, characterized in that: The system comprises: An infrared thermal image and visible light image alignment module, used for synchronously collecting infrared thermal images, visible light images, environmental condition data and operating parameters of typical components in the power transmission line, and aligning the infrared thermal images and visible light images; A thermal radiation data acquisition module, used for acquiring thermal radiation data of typical components in the aligned infrared thermal image through a target detection model; A first comprehensive feature vector generating module, used for fusing the thermal radiation data of each typical component with the environmental condition data to generate a first comprehensive feature vector; An infrared heat potential hazard area identification module for typical components in a transmission line, used to construct a thermal radiation intensity and temperature model according to the first comprehensive feature vector, and identify infrared heat potential hazard areas of typical components in a transmission line through the thermal radiation intensity and temperature model; A second comprehensive feature vector generating module, used for fusing the thermal radiation data of each typical component with the operating parameters to generate a second comprehensive feature vector; A potential fault point identification module for typical components in a transmission line is used to identify potential fault points of typical components in a transmission line according to the second comprehensive feature vector.

13. A computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 11.

Images