High voltage cable joint infrared thermographic time series change analysis processing system

By using a high-voltage cable joint infrared thermal image time-series change analysis and processing system, combined with load current signal and variational mode decomposition algorithm, the false alarm rate and false negative rate problems in high-voltage cable joint thermal anomaly detection have been solved. This system enables accurate qualitative determination and location of heat source type, improving the accuracy and reliability of detection.

CN122416124APending Publication Date: 2026-07-17CHINA THREE GORGES UNIV
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202610548270.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-04-23
Publication Date
2026-07-17

AI Technical Summary

Technical Problem

In existing technologies, thermal anomaly detection methods for high-voltage cable joints have high false alarm and false negative rates under conditions of large load fluctuations, and cannot effectively distinguish thermal anomalies caused by different physical mechanisms, resulting in inaccurate judgments.

Method used

A high-voltage cable joint infrared thermal image time-series change analysis and processing system is adopted. Through the thermal anomaly area determination module, time-series feature extraction module, background separation module, and heat source qualitative determination module, combined with load current signal and variational mode decomposition algorithm, the influence of environmental drift and load fluctuation is removed, the heat source type is identified and the defect is located.

Benefits of technology

It improves the accuracy of thermal anomaly identification, can distinguish between load-driven and self-evolving faults, provides a scientific basis for judgment, accurately locates the physical mechanism and location of internal defects, and avoids false alarms.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122416124A_ABST
    Figure CN122416124A_ABST
Patent Text Reader

Abstract

This invention discloses a system for analyzing and processing the temporal changes of infrared thermal images of high-voltage cable joints, belonging to the field of intelligent diagnostic technology for power equipment. It includes a thermal anomaly area determination module, a temporal feature extraction module, a background separation module, a heat source qualitative determination module, and an output module. Through the collaborative work of these modules, a comprehensive analysis of the thermal images of high-voltage cable joints can be performed, enabling precise location, type identification, and severity classification of cable joint defects. This effectively eliminates interference from environmental factors and load fluctuations, improving the accuracy of heat source determination. It provides a scientific basis for the condition diagnosis of cable joints and strong support for intelligent diagnostics of power equipment, thereby ensuring the safe and stable operation of high-voltage cable lines and reducing the risk of faults.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of intelligent diagnostic technology for power equipment status, and in particular to a system for analyzing and processing the temporal changes of infrared thermal images of high-voltage cable joints. Background Technology

[0002] High-voltage power cables are widely used in urban underground power transmission and distribution networks. Cable joints are critical accessories connecting two cable sections and are also the most complex and fault-prone parts of the entire cable line. During manufacturing or on-site installation, defects such as air gaps, impurities, or inadequate crimping may be introduced into the conductor crimping area inside the joint, leading to higher contact resistance at that location. During current-carrying operation, this high contact resistance generates additional Joule heat, causing the surface temperature of the crimping area to exceed normal levels, forming a localized thermal anomaly. This thermal anomaly initially presents with a very small temperature rise, but slowly increases as the contact resistance deteriorates due to oxidation, creep, loosening, and other factors. If not detected in time, it may eventually lead to insulation breakdown or even joint explosion. Infrared thermal imaging technology can acquire the surface temperature distribution of equipment non-contactly, without interfering with personnel or equipment, making it an important means of detecting thermal anomalies in cable joints. With the decreasing cost of fixed infrared thermal imagers, installing them near cable joints for continuous online monitoring has become feasible in engineering. The time-series thermal image sequence formed by the camera continuously capturing images of the same joint provides a data foundation for analyzing changes in the joint's thermal state over time.

[0003] For example, Chinese invention patent CN112734698B discloses a method and apparatus for diagnosing cable terminal anomalies based on infrared images. The method includes: acquiring an infrared image of the cable terminal to be tested; analyzing the infrared image to obtain the surface temperature of the cable terminal in the infrared image; comparing the surface temperature with a preset temperature threshold to obtain a comparison result; and determining whether the cable terminal to be tested has an anomaly based on the comparison result.

[0004] For example, Chinese invention patent CN118096726B discloses a method for identifying abnormal cable temperatures, which includes: acquiring feature points and first descriptors of feature points in infrared cable images and grayscale cable images; acquiring gradient similarity coefficients of infrared cable images and grayscale cable images; obtaining several feature point clusters; acquiring second descriptors of feature points; performing image matching using the first and second descriptors; identifying temperature abnormal regions in infrared cable images based on the grayscale values ​​of pixels in infrared cable images; and combining the matched feature points to obtain temperature abnormal regions in grayscale cable images.

[0005] However, in the process of implementing the inventive technical solution in the embodiments of this application, it was found that the above-mentioned technology has at least the following technical problems:

[0006] In existing technologies, anomaly detection is mostly achieved by comparing the surface temperature of the cable terminal with a preset threshold. This method is essentially a static criterion model. In practical applications, under conditions of large load fluctuations, normal heating and abnormal heating overlap in temperature amplitude, making it difficult to simultaneously account for false alarm and false negative rates. Furthermore, most existing technologies only utilize the single scalar feature of temperature amplitude, which cannot fully characterize the evolution trend and heat diffusion pattern of thermal anomalies, thus making it difficult to distinguish thermal anomalies caused by different physical mechanisms. Summary of the Invention

[0007] To address the technical problems existing in the prior art, embodiments of the present invention provide a system for analyzing and processing the temporal changes of infrared thermal images of high-voltage cable joints. The technical solution is as follows:

[0008] The high-voltage cable joint infrared thermal image time-series change analysis and processing system includes:

[0009] The thermal anomaly region determination module is used to perform global pixel temperature analysis on thermal images containing thermal anomaly regions to determine the location information of the thermal anomaly regions in the thermal image.

[0010] The temporal feature extraction module is used to receive the continuous temporal thermal image sequence before the thermal image and the synchronously acquired load current signal. Based on the location information and the continuous temporal thermal image sequence, it extracts the temperature time change features and spatial temperature gradient features of the thermal anomaly area. Based on the load current signal, it normalizes the temperature time change features and spatial temperature gradient features to obtain the temporal feature sequence.

[0011] The background separation module is used to remove the global temperature drift component from the time-series feature sequence and output the residual feature sequence.

[0012] The heat source qualitative determination module is used to determine the type of heat source based on the response amplitude of the temperature change rate in the residual feature sequence and the characteristics of the spatial temperature gradient. The heat source type includes at least an internal deep heat source or a surface shallow heat source.

[0013] The output module is used to output diagnostic results based on the location information of the thermal anomaly area and the type of heat source.

[0014] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0015] 1. The high-voltage cable joint infrared thermal image time-series change analysis and processing system provided by this invention, by introducing a physical normalization mechanism for the square of the load current, eliminates the masking effect caused by drastic load fluctuations on the joint temperature rise from the underlying logic, allowing the feature values ​​to directly reflect the essential state of the contact resistance. Combining variational mode decomposition algorithm and time-domain phase alignment operator, the system successfully solves the problem of thermal response hysteresis caused by differences in heat capacity between the target and reference areas, effectively eliminating long-period environmental drift such as day-night temperature cycles. In particular, the geometric shape descriptor established through three-dimensional laser scanning realizes non-uniform environmental sensitivity weighting based on the light-receiving angle and heat dissipation coefficient, completely eliminating false anomalies caused by uneven light reception or reflection on the irregular surface of the joint, greatly improving the confidence level of anomaly identification.

[0016] 2. This invention constructs a qualitative judgment criterion for deep and shallow heat sources by jointly analyzing the residual of temperature change rate and the mean and standard deviation of spatial temperature gradient. The system can sensitively identify shallow interface defects with strong gradients and high stability, and deep insulation loss characteristics with weak gradients and high fluctuations. It also uses a preset physical structure model to map thermal anomaly regions to specific layers such as the press-fit conductor layer, semi-conductive layer, or insulation layer. This in-depth diagnosis, combining qualitative and location analysis, allows maintenance personnel to penetrate the cable sheath and intuitively grasp the physical mechanism and precise location of internal defects.

[0017] 3. This invention addresses the hidden stages of interface degradation evolving into bulk material degradation by introducing first-order difference analysis of current change rate. The system can accurately distinguish between load-driven and self-evolving faults, clearly revealing whether thermal anomalies are caused by load impact or by the inherent deterioration of defects. This in-depth analysis of dynamic operating conditions avoids false alarms during transient processes such as load switching, providing a scientific basis for understanding the evolutionary stages of defects. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0019] Figure 1 A schematic diagram of a high-voltage cable joint infrared thermal image temporal change analysis and processing system module provided in an embodiment of this application.

[0020] Figure 2 The flowchart of the heat source determination and output logic provided in the embodiments of this application is shown. Detailed Implementation

[0021] Embodiments of the present disclosure will now be described in more detail with reference to the accompanying drawings. While some embodiments of the present disclosure are shown in the drawings, it should be understood that embodiments of the present disclosure may be implemented in various forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure.

[0022] It should be understood that the accompanying drawings and embodiments of this disclosure are for illustrative purposes only and are not intended to limit the scope of protection of this disclosure. In the description of the embodiments of this disclosure, the term "comprising" and similar terms should be understood as open-ended inclusion, i.e., "including but not limited to". The term "based on" should be understood as "at least partially based on". The term "one embodiment" or "this embodiment" should be understood as "at least one embodiment". The terms "first", "second", etc., may refer to different or the same objects.

[0023] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0024] Please see Figure 1 The diagram shown is a schematic of the high-voltage cable joint infrared thermal image temporal change analysis and processing system module of the present invention.

[0025] The thermal anomaly region determination module is used to perform global pixel temperature analysis on thermal images containing thermal anomaly regions to determine the location information of the thermal anomaly regions in the thermal image.

[0026] For each frame of thermal image, calculate the mean and standard deviation of the pixel temperature of the entire frame. Pixels with temperature values ​​exceeding the mean of the entire frame plus a preset multiple of the standard deviation are marked as candidate thermal anomaly pixels. Extract all candidate anomaly pixels to establish a candidate anomaly pixel set.

[0027] For example, if the average temperature of a frame is 32°C, the standard deviation is 4°C, and the preset multiplier is 2, then the judgment threshold is 40°C. Pixels with a temperature exceeding 40°C are marked as candidate thermal anomaly pixels, while pixels with a temperature below 40°C are considered background and do not participate in subsequent processing.

[0028] Perform connected component analysis on the candidate pixel set, remove isolated noisy connected components with an area lower than the preset minimum value, and retain connected components with an area that meets the requirements as candidate thermal anomaly regions for the current frame.

[0029] Candidate thermal anomaly pixels may appear in two forms on an image: one is that several pixels are clustered together to form a continuous region (connected region), and the other is that a single pixel or a few pixels are scattered in isolation throughout the image. The former corresponds to a real thermal anomaly region, while the latter is usually a detector dead pixel or noise generated by electromagnetic interference.

[0030] In the temporal dimension, a connected component is identified as a thermal anomaly region only when it appears continuously at similar pixel positions in multiple consecutive frames. Occasional responses in a single frame are not included in the thermal anomaly region. This jointly constrains the determination of thermal anomaly regions in both spatial and temporal dimensions and outputs the pixel range of the thermal anomaly region.

[0031] It should be noted that some candidate regions may be sporadic. For example, if there is a brief external thermal radiation interference (such as a passerby or a brief reflection of sunlight) when a certain frame of image is captured, a candidate thermal anomaly region may be generated in that frame, but there may be no anomaly at that location in the preceding and following frames.

[0032] For example, the connected component near coordinates (47, 83) appears in 10 consecutive frames from frame 101 to frame 110, with a center coordinate offset of no more than 3 pixels, and is determined to be persistent, thus being identified as a thermal anomaly region; while the connected component near coordinates (95, 150) only appears once in frame 105, and does not exist in frames 104 and 106, and is determined to be an occasional response in a single frame, thus being removed.

[0033] The time-series feature extraction module is used to receive a continuous time-series thermal image sequence preceding the thermal image and a synchronously acquired load current signal. Based on the location information and the continuous time-series thermal image sequence, it extracts the temperature-time variation features and spatial temperature gradient features of the thermal anomaly area. Based on the load current signal, it normalizes the temperature-time variation features and spatial temperature gradient features to obtain a time-series feature sequence.

[0034] By performing linear fitting operations on each pixel within the thermal anomaly region across consecutive frames, the temperature rise slope evolving over time is obtained as the rate of temperature change.

[0035] In this embodiment, a calculation window of 30 consecutive frames is used, and the temperature values ​​of the pixel in these 30 frames form a temperature curve that changes over time. A linear fitting operation is performed on this curve, that is, the slope of the linear fitting curve is calculated. A positive slope indicates that the temperature of the pixel continues to rise over time, a negative slope indicates that the temperature decreases, and a slope close to zero indicates that the temperature is stable. The magnitude of the slope directly reflects the speed at which the thermal state of the pixel location evolves over time, that is, the rate of temperature change.

[0036] For example, the pixel at coordinates (3,5) had temperatures of 28.1°C, 28.3°C, 28.4°C...30.1°C over the past 30 frames, showing a linear upward trend with a fitted slope of approximately 0.067°C per frame. Therefore, the temperature change rate at this coordinate point is a temperature increase of 0.067°C per frame. In the same thermal image, the pixel temperature in the background region remained relatively stable at around 28°C, with a slope close to zero, showing a significant difference from the thermally anomalous pixels.

[0037] The temperature difference between a pixel and its eight neighboring pixels is calculated as a spatial temperature gradient characterizing the spatial concentration of heat distribution on the joint surface.

[0038] For each pixel within the thermal anomaly region, the temperature values ​​of that pixel and its eight adjacent pixels (top, bottom, left, right, and four diagonal directions) are read from the current frame's thermal image. The temperature difference between that pixel and its adjacent pixels in each direction is calculated. The magnitude of the temperature difference reflects the spatial concentration of heat at that location; a larger difference indicates a more prominent temperature relative to surrounding pixels, indicating a high concentration of heat; a smaller difference indicates a more even spatial temperature distribution and more uniform heat diffusion. The differences in the eight directions are synthesized to obtain the spatial temperature gradient amplitude and main diffusion direction at that coordinate point. The spatial temperature gradient amplitude is obtained by taking the square root of the sum of the squares of the horizontal and vertical components.

[0039] For example, the temperature at coordinates (4,7) is 42°C, while the temperature of the pixel above it is 38°C, the temperature below it is 39°C, the temperature to its left is 37°C, and the temperature to its right is 36°C. The differences are 4°C, 3°C, 5°C, and 6°C, respectively. The resulting gradient amplitude is large, indicating that this point is a thermal core with highly concentrated local heat, corresponding to the typical spatial characteristics of a shallow heat source.

[0040] In another example, the temperature at coordinates (8,9) is 34°C, while the temperatures of the surrounding pixels are between 32°C and 33°C. The difference in each direction is between 1 and 2°C. The gradient amplitude is small and the direction is dispersed, which corresponds to the heat dissipation spot characteristics formed on the surface after the deep heat source diffuses through the insulating medium.

[0041] A two-dimensional coordinate matrix is ​​established for the thermal anomaly region. Operator operations are performed synchronously on the temperature change rate and spatial temperature gradient of each coordinate point in the two-dimensional coordinate matrix to obtain the time-series feature matrix.

[0042] Operator operations are performed simultaneously on the rate of temperature change and the spatial temperature gradient at each coordinate point in the two-dimensional coordinate matrix, specifically including:

[0043] Using the pixel coordinate system of the thermal image as a reference, a two-dimensional coordinate matrix is ​​established for all pixels in the thermal anomaly area. Each coordinate point in the two-dimensional coordinate matrix corresponds to the spatial position of a pixel.

[0044] For each coordinate point in the two-dimensional coordinate matrix, the temperature sequence of that coordinate point in a continuous preset number of thermal images is taken along the time axis. A linear slope is fitted to the temperature sequence within a sliding time window to obtain the temperature change rate value of that coordinate point.

[0045] Simultaneously, a spatial gradient operator is executed on the current frame thermal image within the thermal anomaly region to calculate the gradient components of each coordinate point in the horizontal and vertical directions, and synthesize the spatial temperature gradient magnitude and gradient direction of that coordinate point.

[0046] Repeat the above operation for all coordinate points within the thermal anomaly region, and obtain the time series feature matrix by arranging them according to their coordinate positions.

[0047] The temporal feature matrix preserves the complete spatial structure information of the thermal anomaly region: each position in the matrix records not the original temperature value, but the evolution rate of the pixel at that position in the time dimension and the gradient intensity in the spatial dimension. The two types of information are superimposed in the same spatial coordinate system, providing a feature description with both time and spatial dimensions for subsequent analysis.

[0048] For example, a coordinate point (3,5) in a thermal anomaly region.

[0049] Simultaneously execute the following on coordinate point (3,5):

[0050] Take the temperature values ​​at (3,5) over the past 30 frames along the time axis, arrange them into a column, and calculate the slope of this column to obtain the rate of temperature change at that coordinate point.

[0051] Along the spatial axis in the current frame of the image, based on the temperature difference between the point (3,5) and its adjacent pixels above, below, left, and right, calculate the magnitude and direction of the difference to obtain the spatial temperature gradient of that coordinate point.

[0052] The rate of temperature change and spatial temperature gradient of all coordinate points within the thermal anomaly region are arranged according to the original spatial coordinates, forming a feature matrix corresponding to the spatial structure of the thermal anomaly region.

[0053] The synchronously acquired load current signal is introduced into the time-series feature matrix as a physical constraint. For the normalized matrix corresponding to each frame image, the spatial statistical feature extraction operator is used to calculate the single-frame feature vector representing the overall heating characteristics of the abnormal area. The single-frame feature vectors are then concatenated in time-series logic according to the timestamp order and encapsulated to form a time-series feature sequence.

[0054] It should be noted that the temperature change rate and spatial temperature gradient in the time-series feature matrix are both affected by the magnitude of the load current. When the load current increases, even if the joint has no defects, the normal Joule heat generated by the contact resistance will cause the temperature change rate and gradient values ​​to be larger. To eliminate the interference of load current fluctuations on the feature values, the feature value of each coordinate point in the time-series feature matrix is ​​normalized by dividing it by the square of the load current at the corresponding moment in the same frame. This ensures that the feature values ​​only reflect the thermal anomalies caused by the defect itself, and are not affected by normal load changes.

[0055] After normalization, for the temporal feature matrix corresponding to each frame of thermal image, the spatial statistical feature extraction operator is used to statistically summarize the normalized feature values ​​of all coordinate points in the matrix. This calculates statistical quantities such as the average temperature change rate, average gradient magnitude, and gradient concentration of the overall thermal anomaly region. These statistical quantities are then combined into a single-frame feature vector representing the overall thermal characteristics of the thermal anomaly region in that frame. The single-frame feature vector compresses information spatially distributed across multiple pixels into a set of values ​​describing the overall characteristics, facilitating subsequent temporal analysis and processing.

[0056] The above operation is repeated for each frame of continuously acquired thermal image, generating a single-frame feature vector for each frame, with each vector accompanied by a corresponding acquisition timestamp. The single-frame feature vectors of each frame are arranged sequentially according to their timestamps to form a feature sequence with time as the axis, namely the time-series feature sequence.

[0057] The temporal feature sequence records the evolution trajectory of the overall heating characteristics of the thermal anomaly region over a period of time. This sequence serves as the input for both the subsequent background separation module and the heat source qualitative determination module. By analyzing the changes in feature values ​​over time within the sequence, the defect development trend can be identified, and the depth type of the heat source can be determined.

[0058] The background separation module is used to remove the global temperature drift component from the time-series feature sequence and output the residual feature sequence.

[0059] Within the infrared field of view, a healthy reference region with the same material as the thermal anomaly region is automatically calibrated, and the difference in time constant between the two regions in response to environmental thermal excitation is calculated. By constructing a time-domain phase alignment operator, inertial compensation and phase shift are performed on the time-series feature sequence.

[0060] The healthy baseline region refers to the area within the thermal imaging field of view that is made of the same material as the thermal anomaly region but is not affected by internal heat sources in the joint. Ambient temperature changes affect both the thermal anomaly region and the healthy baseline region simultaneously. By comparing the differences in their temperature responses, temperature changes caused by environmental factors can be isolated from the temperature sequence of the thermal anomaly region.

[0061] During the system initialization phase (when the connector is newly installed and defects have not yet formed), all pixels within the thermal image field of view are clustered by material category. The set of pixels with the same material as the crimped sleeve and whose long-term temperature variance is below a threshold is identified as the healthy baseline region. Alternatively, during the operation phase, a pixel region on the same cable sheath that is sufficiently far from the connector, has a high correlation between its temperature history curve and the thermal anomaly region, but whose absolute value does not deviate continuously, is selected as the baseline. Calibration is prioritized during the initialization phase; if calibration cannot be completed during initialization, it is performed during the operation phase.

[0062] For example, within the field of view of a thermal image, the pixels on the surface of the crimp sleeve of the connector constitute candidates for thermal anomaly areas. The pixel temperature on the sheath section of the same material about 15 cm away from the connector is stable over a long period of time and is highly consistent with the temperature change trend of the crimp sleeve area. This area is then designated as the healthy baseline area.

[0063] The time constant difference refers to the time lag in the temperature response of the thermal anomaly region and the reference region to the same environmental thermal excitation. This is because the two regions have different heat capacities. The metal region of the crimped sleeve has a small heat capacity and responds quickly, while the insulating sheath has a large heat capacity and responds slowly. Therefore, the time constants of the temperature rise curves of the two regions are different under the same ambient temperature step.

[0064] The method for extracting the time constant utilizes the natural periodic thermal excitation of the diurnal temperature cycle. Daytime temperatures rise and nighttime temperatures fall, exhibiting a regular periodic change in ambient temperature. Both the thermal anomaly region and the healthy baseline region respond to this change, but at different rates. An exponential function is fitted to the temperature history sequences of each region to extract their respective thermal response time constants. The difference between the two is calculated to obtain the time lag of the thermal anomaly region relative to the healthy baseline region. For example, the healthy baseline region reaches a new steady-state temperature approximately 8 minutes after a step change in ambient temperature, while the thermal anomaly region, due to its larger heat capacity, takes approximately 20 minutes to reach steady state. The difference in their time constants is approximately 12 minutes, which represents the phase shift that needs to be compensated for.

[0065] The construction process of the phase alignment operator includes:

[0066] Shifting the reference sequence by this offset on the time axis is equivalent to performing a fractional delay filter on the reference sequence, aligning the phase of the reference sequence with the thermal anomaly region before performing the difference, thus eliminating spurious peaks caused by differences in thermal capacity.

[0067] Continuing the previous example, the healthy baseline sequence is shifted forward by 12 minutes on the time axis to align the peak temperature response of the baseline sequence with the peak temperature response of the thermal anomaly region in time. Since 12 minutes is not necessarily an integer frame interval, the shift operation is achieved through fractional delay filtering, i.e., interpolation is performed between adjacent frames to generate the baseline sequence value corresponding to the non-integer frame position. After phase alignment, the aligned baseline sequence is subtracted from the temperature sequence of the thermal anomaly region. At this point, the two sequences are synchronized in time, and no spurious peaks due to phase misalignment will appear in the subtraction result.

[0068] If phase alignment is not performed and subtraction is performed directly, when the ambient temperature changes rapidly, the response of the thermal anomaly region and the response of the reference region will be out of sync in time. The subtraction result will show alternating positive and negative spikes. These spikes are not real defect signals, but pseudo-differential responses caused by phase misalignment, which will interfere with subsequent defect determination.

[0069] Even after phase alignment, the sequence still contains two overlapping components: one is a low-frequency trend component caused by long-period changes in ambient temperature (diurnal cycle, seasonal variation), and the other is a mid-frequency disturbance component caused by the evolution of heat sources from internal defects in the joint. These two components are difficult to separate directly in the time domain because they coexist and superimpose within the same time period.

[0070] The variational mode decomposition algorithm is used to decompose the time-series characteristic sequence of the thermal anomaly region into a trend component reflecting the long-term evolution of the environment and a disturbance component reflecting the internal intrinsic loss by subtracting the time-series characteristic sequence from the phase-aligned reference sequence. This achieves frequency domain decoupling between environmental noise and fault signals.

[0071] By combining the three-dimensional geometric shape descriptor of the joint surface, non-uniform environmental sensitivity weights are assigned according to the light angle of each pixel coordinate point relative to the ambient radiation source and the surface heat dissipation coefficient. After the trend component is weighted according to the environmental sensitivity weight matrix, it is deducted pixel by pixel from the time-series feature sequence, thereby outputting a residual feature sequence that can accurately characterize the evolution trend of physical defects inside the joint.

[0072] The three-dimensional geometric shape descriptor of the connector surface, specifically obtained through the following process:

[0073] The surface of a cable joint is not flat, but rather composed of multiple sections of rotation with varying curvatures, including a crimped sleeve, insulation layer, and heat shrink tubing. The degree of absorption of environmental radiation (such as solar radiation and indoor lighting) and the convective heat dissipation capacity at different locations on the joint surface vary due to differences in orientation and curvature. Subtracting the trend component from the time-series feature sequence with uniform weights would introduce errors. During system installation, a 3D laser scanner is used to perform a full-circumference scan of the cable joint surface, acquiring 3D point cloud data. The point cloud data is then denoised and fitted to surfaces to extract the spatial coordinates, normal vector, and local curvature of each sampling point on the joint surface.

[0074] The three-dimensional point cloud coordinate system is calibrated with the image coordinate system of the fixed infrared thermal imager to establish the mapping relationship between the sampling points on the three-dimensional surface and the pixel coordinates of the thermal image.

[0075] By mapping the normal vector and curvature value of each pixel in the thermal image onto the thermal image plane, each pixel in the thermal image is assigned a corresponding normal vector and curvature value, thus forming a three-dimensional geometric shape descriptor.

[0076] The light-receiving angle refers to the cosine of the angle between the surface normal vector of the pixel and the direction vector of the ambient radiation source. The smaller the angle (facing the radiation source), the greater the influence of ambient radiation on the pixel, and the higher the environmental sensitivity weight should be assigned. The heat dissipation coefficient refers to the surface convection heat dissipation capacity of the pixel location, which depends on the surface curvature and orientation. The higher the heat dissipation coefficient, the faster the response to changes in ambient temperature, and the higher the corresponding environmental sensitivity weight.

[0077] The calculation method for the light-receiving angle weight is as follows: the cosine value of the angle between the surface normal vector of a pixel and the direction vector of the ambient radiation source is the light intensity index of that pixel. When the normal vector is directly facing the radiation source, the angle is zero degrees and the cosine value is one, and the pixel is most affected by ambient radiation; when the normal vector is perpendicular to the direction of the radiation source, the cosine value is zero, and the pixel is almost unaffected by radiation in that direction.

[0078] The heat dissipation coefficient weighting is calculated as follows: Protruding areas with high curvature have a relatively large contact area with the air, resulting in strong convective heat dissipation and a rapid response to changes in ambient temperature; flat areas with low curvature dissipate heat relatively slowly. The heat dissipation coefficient is estimated based on the curvature; the greater the curvature, the higher the heat dissipation coefficient.

[0079] The cosine of the light-receiving angle is multiplied by the heat dissipation coefficient to obtain the original environmental sensitivity weight for each pixel. The original environmental sensitivity weights of all pixels within the thermal anomaly region are normalized so that the sum of the environmental sensitivity weights is one, resulting in an environmental sensitivity weight matrix. This environmental sensitivity weight matrix is ​​then used to weight the trend component pixel-by-pixel. Each pixel has a corresponding proportion of the trend component deducted based on its actual degree of environmental influence, rather than deducting the same value uniformly for all pixels. This accurately eliminates the environmental response differences caused by the non-uniform geometry of the joint surface, and the output residual feature sequence retains only the true thermal change signal caused by the evolution of physical defects inside the joint.

[0080] Please see Figure 2 The diagram shown is a flowchart of the heat source determination output logic provided in an embodiment of this application. It describes the cable thermal anomaly diagnosis decision process based on residual feature sequences.

[0081] The heat source qualitative determination module is used to determine the type of heat source based on the response amplitude of the temperature change rate in the residual feature sequence and the characteristics of the spatial temperature gradient. The heat source type includes at least an internal deep heat source or a surface shallow heat source.

[0082] In the residual feature sequence, each time point corresponds to a temperature change rate residual value, reflecting the actual thermal evolution rate of the thermal anomaly region after removing background interference at that moment. The period with a large absolute value of the temperature change rate residual indicates that the heat source inside the joint has generated significant thermal changes during that period, which is the interval where the defect signal is most concentrated, and is called the temperature change rate response peak interval.

[0083] Identify the peak interval of the temperature change rate response in the residual feature sequence. The peak interval of the temperature change rate response is a continuous time window that satisfies that the absolute value of the temperature change rate residual is greater than or equal to a preset proportional threshold.

[0084] The calibration method for the preset proportional threshold is as follows: During the system initialization phase, a residual feature sequence of the joint during a sufficiently long period of healthy operation is collected, and the distribution of the absolute value of the temperature change rate residual in this sequence is statistically analyzed. The 95th percentile of the absolute value of the temperature change rate residual is used as the reference benchmark for the preset proportional threshold. Periods exceeding this benchmark are considered to have thermal responses exceeding the normal fluctuation range. In actual engineering, this threshold is usually further corrected through field calibration experiments to balance the false alarm rate and the missed alarm rate. For example, if the absolute value of the temperature change rate residual during healthy operation is usually within 0.02°C per frame, the 95th percentile of 0.03°C is used as the threshold.

[0085] Calculate the mean and standard deviation of the spatial temperature gradient residuals within the peak interval of the rate of temperature change response. The mean reflects the overall intensity of the spatial temperature gradient in the thermal anomaly region within the peak interval, while the standard deviation reflects the temporal stability of the gradient. A high mean of the spatial temperature gradient residuals indicates that heat is highly concentrated at a local location, and the gradient field spatially manifests as a sharp thermal core; a low mean indicates that heat diffuses evenly to the surrounding area, and the gradient field spatially manifests as a gently spreading heat patch. A low standard deviation indicates that the gradient remains stable across frames within the peak interval, and the location and intensity of the heat source remain essentially unchanged; a high standard deviation indicates that the gradient fluctuates significantly across frames, and the intensity or location of the heat source changes significantly over time.

[0086] If the mean of the spatial temperature gradient residual is greater than or equal to the high threshold and the standard deviation of the spatial temperature gradient residual is less than or equal to the low threshold, it is determined to be a shallow surface heat source.

[0087] If the mean of the spatial temperature gradient residual is less than the low threshold and the standard deviation of the spatial temperature gradient residual is greater than the high threshold, it is determined to be an internal deep heat source.

[0088] The shallow surface heat source characterizes the rapid thermal response caused by interface defects such as increased contact resistance and oxidation corrosion. Its spatial gradient is strong and stable. After the heat is generated, it can reach the surface without being conducted through a thick insulating medium. Therefore, the surface gradient field is strong and concentrated, the thermal core position is fixed, and the gradient value of each frame is stable.

[0089] The deep internal heat source characterizes the slow heat accumulation caused by insulation aging, partial discharge, or dielectric loss, with a weak spatial gradient and significant fluctuations. Heat diffuses and attenuates through a thick insulating medium before reaching the surface, resulting in a weak and diffuse surface gradient field. Furthermore, due to the influence of various factors on the heat conduction path within the medium, the surface gradient fluctuates considerably between frames.

[0090] The high and low thresholds are determined jointly through simulation and laboratory calibration. Cable joint samples with known defect depths are prepared in the laboratory to simulate two types of defects: poor contact (shallow) and insulation aging (deep). Time-series thermal images are acquired under controlled conditions, and the mean and standard deviation of the spatial temperature gradient residuals within the peak intervals of each sample are extracted. The boundary interval between the two types of sample data distributions is used as the basis for setting the high and low thresholds. Typically, the high threshold corresponds to the 10th percentile of the mean of the shallow sample in both distributions, and the low threshold corresponds to the 90th percentile of the mean of the deep sample. The interval between the two thresholds is the transition state determination range.

[0091] Heat source types also include transition state heat sources. The criteria for determining a transition state heat source are:

[0092] The mean of the spatial temperature gradient residual is between the low threshold and the high threshold, and the standard deviation of the spatial temperature gradient residual is between the low threshold and the high threshold. At this time, the heat source is in the intermediate stage of the evolution from interface degradation to material degradation.

[0093] Furthermore, considering the spatial location of the thermal anomaly region, if the thermal anomaly region is located in a structurally sensitive area, it is further identified as a transitional heat source.

[0094] When the mean and standard deviation of the spatial temperature gradient residual both fall in the middle range between the high and low thresholds, the spatial characteristics of the heat source do not possess the strong gradient stability characteristics of shallow heat sources, which have higher heat dissipation coefficients, faster responses to changes in ambient temperature, and corresponding higher environmental sensitivity weights, nor do they possess the weak gradient high fluctuation characteristics of deep heat sources. This indicates that the heat source is in the intermediate stage of the evolution from interface defects to bulk material defects, that is, poor contact or interface deterioration on the joint surface has begun to extend into the interior of the insulation layer, but has not yet fully developed into a typical bulk material loss type defect.

[0095] Structurally sensitive areas are regions within cable joints where electric field or thermal stress concentration occurs due to abrupt structural changes. Typical locations include the ends of stress cones and shielding overlaps. These locations are inherently weak points in the joint design; electric field distortion and abrupt thermal stress occur at these points, making it easier for defects to evolve from interface degradation to bulk material degradation. The spatial location of structurally sensitive areas is pre-determined and stored in the system based on the joint type's design drawings during system installation.

[0096] Calculate the first difference of the load current sequence and normalize it to the rate of change of current, reflecting the relative magnitude of current change between adjacent time points. Statistically calculate the absolute value of the mean of the rate of change of current within the peak interval of the temperature change rate response to determine whether the load current has changed significantly during the period of this peak interval.

[0097] If the absolute value of the mean of the rate of change of current within the peak range of the rate of change of temperature response is greater than or equal to the high rate of change threshold, it indicates that the load current is changing rapidly when the peak range appears, and the occurrence of thermal anomalies is highly synchronized with load fluctuations. The current excitation of this transitional heat source is caused by load changes, and the mean of the spatial temperature gradient residual is between the low threshold and the high threshold. In this case, it is determined to be a load-driven transitional heat source. The thermal anomaly of this type of heat source may weaken after the load recovers and stabilizes, but the interface degradation at the structural level still exists and needs to be continuously monitored.

[0098] If the absolute value of the mean rate of change of current is less than the low rate of change threshold, and the mean value of the residual of spatial temperature gradient is between the low threshold and the high threshold, it indicates that the load current is basically stable when the peak interval occurs, and the occurrence of thermal anomalies is unrelated to load changes. It is an endogenous thermal evolution caused by the development of defects themselves. It is then judged as a self-evolving transitional heat source. This type of heat source continues to develop under the condition of constant load, and its danger level is higher than that of load-driven types. It should be prioritized for maintenance.

[0099] The calibration methods for high and low rate of change thresholds include: statistically analyzing the daily fluctuation pattern of the load current during normal operation of the line, taking the 90th percentile of the current change rate within the normal fluctuation range as the low rate of change threshold, and taking the 10th percentile of the current change rate corresponding to the typical large current change caused by operations such as switching large loads and line switching as the high rate of change threshold. The interval between the two corresponds to an intermediate state where the cause cannot be clearly attributed. At this time, the system outputs an uncertain label and triggers manual review.

[0100] The output module is used to output diagnostic results based on the location information of the thermal anomaly area and the type of heat source.

[0101] Based on the physical structure model of the connector pre-established during system installation, the pixel coordinate range of the thermal anomaly area is mapped to the physical structure coordinate system of the cable connector to determine the specific structural layers inside the connector corresponding to the thermal anomaly area, including the crimped conductor layer, the semiconducting layer, and the insulation layer.

[0102] The establishment of the mapping relationship relies on the pre-entered physical structure model of the connector during the system installation phase. This model records the geometric boundaries of each structural layer of the connector (crimped conductor layer, semiconducting layer, and insulating layer) in three-dimensional space, as well as the corresponding projection positions of these boundaries in the thermal image pixel coordinate system. By using the established three-dimensional geometric shape descriptor and coordinate system calibration relationship, the pixel coordinate range of the thermal anomaly area is converted into the corresponding position in the connector physical structure coordinate system, determining which structural layer's geometric boundary range this position falls within, thereby identifying the structural layer position.

[0103] Based on the combination of heat source type label and structural layer, a preset defect type knowledge base is matched to output a defect type description, including at least one of poor contact, insulation aging, interface degradation, and transitional heat source.

[0104] The defect type knowledge base is a pre-built rule table indexed by combinations of heat source type tags and structural layers, outputting corresponding defect type descriptions. The knowledge base content is based on the physical laws governing cable joint defect mechanisms, reflecting typical defect forms corresponding to different combinations of heat source types and structural layers in engineering practice.

[0105] The matching logic is as follows: the heat source type label output by the heat source qualitative judgment module and the structural layer determined in the current step are used together as query conditions to retrieve the corresponding entries in the knowledge base and output the defect type description.

[0106] Typical matching relationships include:

[0107] The combination of a shallow surface heat source and a press-fit conductor layer corresponds to the defect type of poor contact. The physical mechanism is that oxidation of the conductor contact surface in the press-fit area or incomplete press-fitting leads to increased contact resistance. The combination of a deep internal heat source and an insulating layer corresponds to the defect type of insulation aging. The physical mechanism is that the dielectric loss of the insulating medium increases under the long-term action of electric and thermal fields. The combination of a transitional heat source and a semiconductive layer corresponds to the defect type of interface degradation. The physical mechanism is that the interface between the semiconductive layer and the insulating layer undergoes local peeling or the formation of conductive channels in areas of concentrated electric field. When the transitional heat source is located in a structurally sensitive area, the corresponding defect type description includes a staged description of the evolution from interface degradation to bulk material degradation.

[0108] Based on the current amplitude and rate of change of the heat source type label and residual feature sequence, the severity of the defect is classified and the defect severity level is output, specifically including:

[0109] The mean residual amplitude and residual change rate of the thermal anomaly region within the current time window are extracted from the residual feature sequence. The mean residual amplitude is used to reflect the degree of abnormal heat accumulation caused by the current defect, and the residual change rate is used to reflect the speed of defect development.

[0110] Based on the combination of heat source type label, mean residual amplitude, and residual change rate, the defect severity level is output according to the following rules:

[0111] When the average residual amplitude is lower than the first threshold and the residual change rate is lower than the first rate threshold, it is judged as a minor defect, indicating that the defect accumulation is small and the development is slow, and it is in the nascent stage. It has not yet posed a substantial threat to the normal operation of the joint, and it is recommended to include it in regular tracking and monitoring.

[0112] When the mean residual amplitude is between the first threshold and the second threshold, or the residual change rate is between the first rate threshold and the second rate threshold, it is judged as a moderate defect. This indicates that the accumulation of defects has reached a certain level or the development speed has accelerated significantly, and it is in the development stage. It is recommended to shorten the detection cycle and include it in key tracking. Power outage maintenance should be arranged according to the development trend.

[0113] When the average residual amplitude exceeds the second threshold, or the residual change rate exceeds the second rate threshold, it is judged as a severe defect, indicating that the accumulated defect has reached a dangerous level or is rapidly deteriorating and is in a critical stage. Power outage maintenance must be arranged immediately, and operation with defects must not continue.

[0114] The determination of the first and second thresholds is divided into two stages: simulation calibration and on-site correction. In the simulation calibration stage, finite element thermal simulation is performed on joints with different defect severity levels to generate residual amplitude distribution ranges corresponding to mild, moderate, and severe defects. The residual amplitude corresponding to the boundary between mild and moderate defects is set as the first threshold, and the residual amplitude corresponding to the boundary between moderate and severe defects is set as the second threshold. In the on-site correction stage, based on existing power outage maintenance records, the actual severity of defects discovered during disassembly and inspection is compared with the residual amplitude records during system operation. The simulation calibration values ​​are then corrected to ensure that the consistency rate between the classification results and the actual defect severity meets engineering requirements. The determination of the first and second rate thresholds is similar, involving statistical analysis of typical intervals of residual change rates during the development of defects of different severity levels, combined with power outage maintenance data for correction. In practical engineering, for shallow heat sources, the first rate threshold usually corresponds to a cumulative increase of 0.5°C in the residual change rate over 24 consecutive hours, and the second rate threshold usually corresponds to a cumulative increase of 2°C over 24 consecutive hours; for deep heat sources, the corresponding threshold is usually relaxed to the cumulative increase over 72 consecutive hours.

[0115] The spatial location, structural layer, defect type description, and defect severity level of the thermal anomaly area are integrated into a structured diagnostic report and output to the monitoring terminal. This allows maintenance personnel to obtain complete defect location and diagnostic information without touching the equipment, and to develop targeted repair plans accordingly.

[0116] Through the above description of the embodiments, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the above functions can be divided into different functional modules to complete all or part of the functions described above.

[0117] In the embodiments provided in this application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are merely illustrative; for instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another device, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection between devices or units through some interfaces, and may be electrical, mechanical, or other forms.

[0118] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A system for analyzing and processing the temporal changes of infrared thermal images of high-voltage cable joints, characterized in that, include: The thermal anomaly region determination module is used to perform global pixel temperature analysis on thermal images containing thermal anomaly regions to determine the location information of the thermal anomaly regions in the thermal images. The time-series feature extraction module is used to receive a continuous time-series thermal image sequence preceding the thermal image and a synchronously acquired load current signal, extract the temperature-time variation features and spatial temperature gradient features of the thermal anomaly region based on the location information and the continuous time-series thermal image sequence, and normalize the temperature-time variation features and spatial temperature gradient features based on the load current signal to obtain a time-series feature sequence. The background separation module is used to remove the global temperature drift component from the time-series feature sequence and output the residual feature sequence; The heat source qualitative determination module is used to determine the type of heat source based on the response amplitude of the temperature change rate and the characteristics of the spatial temperature gradient in the residual feature sequence. The heat source type includes at least an internal deep heat source or a surface shallow heat source. The output module is used to output diagnostic results based on the location information of the thermal anomaly area and the type of heat source.

2. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 1, characterized in that: The determination of the location information of the thermal anomaly region in the thermal image specifically includes: For each frame of thermal image, calculate the mean and standard deviation of the pixel temperature of the entire frame. Pixels with temperature values ​​exceeding the mean of the entire frame plus a preset multiple of the standard deviation are marked as candidate thermal anomaly pixels. Extract all candidate anomaly pixels to establish a candidate anomaly pixel set. Perform connected component analysis on the candidate pixel set, remove isolated noisy connected components with an area lower than the preset minimum value, and retain connected components with an area that meets the requirements as candidate thermal anomaly regions for the current frame; In the temporal dimension, a connected component is identified as a thermal anomaly region only when it appears continuously at similar pixel positions in multiple consecutive frames. Occasional responses in a single frame are not included in the thermal anomaly region, and the pixel range of the thermal anomaly region is output.

3. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 1, characterized in that: The obtained time-series feature sequence specifically includes: By performing linear fitting operations on each pixel in the thermal anomaly region across consecutive frames, the temperature rise slope evolving over time is obtained as the rate of temperature change. The temperature difference between a pixel and its eight neighboring pixels is calculated as a spatial temperature gradient characterizing the spatial concentration of heat distribution on the joint surface. A two-dimensional coordinate matrix is ​​established for the thermal anomaly region. Operator operations are performed simultaneously on the temperature change rate and spatial temperature gradient of each coordinate point in the two-dimensional coordinate matrix to obtain the time-series feature matrix. The synchronously acquired load current signal is introduced into the time-series feature matrix as a physical constraint. For the normalized matrix corresponding to each frame image, the spatial statistical feature extraction operator is used to calculate the single-frame feature vector. The single-frame feature vectors are then concatenated according to the timestamp order and encapsulated to form a time-series feature sequence.

4. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 3, characterized in that: The simultaneous execution of operator operations on the temperature change rate and spatial temperature gradient at each coordinate point in the two-dimensional coordinate matrix specifically includes: Based on the pixel coordinate system of the thermal image, a two-dimensional coordinate matrix is ​​established for all pixels in the thermal anomaly area. Each coordinate point in the two-dimensional coordinate matrix corresponds to the spatial position of a pixel. For each coordinate point in the two-dimensional coordinate matrix, the temperature sequence of that coordinate point in a continuous preset number of thermal images is taken along the time axis. A linear slope is fitted to the temperature sequence within a sliding time window to obtain the temperature change rate value of that coordinate point. Simultaneously, the spatial gradient operator is executed on the current frame thermal image within the thermal anomaly region to calculate the gradient components of each coordinate point in the horizontal and vertical directions, and synthesize the spatial temperature gradient magnitude and gradient direction of that coordinate point. Repeat the above operation for all coordinate points within the thermal anomaly region, and obtain the time series feature matrix by arranging them according to their coordinate positions.

5. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 1, characterized in that: The step of removing the global temperature drift component from the time-series feature sequence specifically includes: Within the infrared field of view, a healthy reference region with the same material as the thermal anomaly region is automatically calibrated, and the difference in time constant between the two regions in response to environmental thermal excitation is calculated. By constructing a temporal phase alignment operator, inertial compensation and phase shift are performed on the temporal feature sequence. The variational mode decomposition algorithm is used to decompose the phase-aligned time series feature sequence into trend components and disturbance components; Combining the three-dimensional geometric shape descriptor of the connector surface, non-uniform environmental sensitivity weights are assigned according to the light angle of each pixel coordinate point relative to the ambient radiation source and the surface heat dissipation coefficient. The trend component is non-linearly weighted and subtracted from the temporal feature sequence to output the residual feature sequence.

6. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 5, characterized in that: The specific process for obtaining the three-dimensional geometric shape descriptor of the connector surface includes: The three-dimensional point cloud data of the joint surface is acquired, and the point cloud data is denoised and fitted to a surface. The spatial coordinates, normal vector and local curvature of each sampling point on the joint surface are extracted. The three-dimensional point cloud coordinate system is calibrated with the image coordinate system of the fixed infrared thermal imager to establish the mapping relationship between the three-dimensional surface sampling points and the pixel coordinates of the thermal image; By mapping the normal vector and curvature value of each pixel in the thermal image onto the thermal image plane, each pixel in the thermal image is assigned a corresponding normal vector and curvature value, thus forming a three-dimensional geometric shape descriptor.

7. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 1, characterized in that: The specific process for determining the type of heat source includes: In the residual feature sequence, identify the peak interval of the temperature change rate response, wherein the peak interval of the temperature change rate response is a continuous time window that satisfies that the absolute value of the temperature change rate residual is greater than or equal to a preset proportional threshold. Calculate the mean and standard deviation of the spatial temperature gradient residuals within the peak interval of the rate of temperature change response; If the mean of the spatial temperature gradient residual is greater than or equal to the high threshold and the standard deviation of the spatial temperature gradient residual is less than or equal to the low threshold, it is determined to be a shallow surface heat source. If the mean of the spatial temperature gradient residual is less than the low threshold and the standard deviation of the spatial temperature gradient residual is greater than the high threshold, it is determined to be an internal deep heat source.

8. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 7, characterized in that: The heat source type also includes a transition state heat source, and the determination criteria for the transition state heat source are as follows: The mean of the spatial temperature gradient residual is between the low threshold and the high threshold, and the standard deviation of the spatial temperature gradient residual is between the low threshold and the high threshold. At this time, the heat source is in the intermediate stage of the evolution from interface degradation to material degradation. Furthermore, considering the spatial location of the thermal anomaly region, if the thermal anomaly region is located in a structurally sensitive area, it is further identified as a transitional heat source. Calculate the first difference of the load current sequence and normalize it to the rate of change of current; If the absolute value of the mean of the rate of change of current within the peak range of the rate of change of temperature response is greater than or equal to the high rate of change threshold, and the mean of the spatial temperature gradient residual is between the low threshold and the high threshold, then it is determined to be a load-driven transitional heat source. If the absolute value of the mean rate of change of current is less than the low rate of change threshold, and the mean value of the spatial temperature gradient residual is between the low threshold and the high threshold, then it is determined to be a self-evolving transitional heat source.

9. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 1, characterized in that: The process of outputting diagnostic results based on the location information of the thermal anomaly area and the type of heat source includes: Based on the pre-established physical structure model of the joint, the pixel coordinate range of the thermal anomaly area is mapped to the physical structure coordinate system of the cable joint to determine the specific structural layers inside the joint corresponding to the thermal anomaly area, including the crimped conductor layer, the semiconductive layer and the insulation layer. Based on the combination of heat source type label and structural layer, a preset defect type knowledge base is matched to output a defect type description, including at least one of poor contact, insulation aging, interface degradation and transitional heat source; Based on the current amplitude and rate of change of the heat source type label and residual feature sequence, the severity of the defect is classified and the defect severity level is output. The spatial location, structural layer, defect type description, and defect severity level of the thermal anomaly area are integrated into a structured diagnostic report and output to the monitoring terminal.

10. The high-voltage cable joint infrared thermal image time-series change analysis and processing system as described in claim 9, characterized in that: The specific process for determining the severity level of the output defect includes: Extract the mean residual amplitude and residual change rate of the thermal anomaly region within the current time window from the residual feature sequence. The mean residual amplitude is used to reflect the degree of abnormal heat accumulation generated by the current defect, and the residual change rate is used to reflect the speed of defect development. Based on the combination of heat source type label, mean residual amplitude, and residual change rate, the defect severity level is output according to the following rules: When the mean residual amplitude is lower than the first threshold and the residual change rate is lower than the first rate threshold, it is judged as a minor defect. When the mean residual amplitude is between the first threshold and the second threshold, or when the residual change rate is between the first rate threshold and the second rate threshold, it is judged as a moderate defect. When the average residual amplitude exceeds the second threshold, or the residual change rate exceeds the second rate threshold, it is judged as a severe defect.

Citation Information

Patent Citations

  • Cable terminal abnormality diagnosis method and device based on infrared image

    CN112734698B

  • A method for identifying abnormal cable temperature

    CN118096726B