Electric energy metering box with temperature monitoring function and monitoring method

Abstract

The present application relates to a kind of electric energy metering box with temperature monitoring function and monitoring method, it is related to electric energy metering box technical field, including, to the infrared thermal imaging scanning of electric energy metering box incoming and outgoing line terminal, obtains temperature distribution chart;When finding abnormal high temperature spot, trigger system real-time acquisition three-phase current data;Based on current data, carry out load current analysis, generate load characteristic curve, and extract its fluctuation characteristics, obtain load fluctuation parameter;Combining load fluctuation parameter evaluation equipment operating state, determine its health condition, and accordingly execute corresponding operation and maintenance strategy, the technical problem that the temperature monitoring and load analysis as independent monitoring means in prior art usually, cannot establish the correlation between the two, lead to when discovering temperature anomaly, it is difficult to accurately judge is equipment itself fault or the normal temperature rise caused by load characteristic, in turn influence the accuracy of fault diagnosis and the timeliness of operation and maintenance decision.

Description

Technical Field

[0001] This invention relates to the field of electricity metering box technology, and in particular to an electricity metering box with temperature monitoring function and a monitoring method thereof. Background Technology

[0002] In the operation and maintenance of power systems, electricity metering boxes, as key equipment for electricity metering and distribution, play a crucial role in ensuring power supply security through monitoring their operational status. Current technologies primarily employ periodic inspections and single-parameter monitoring for monitoring electricity metering boxes. Specifically, maintenance personnel use infrared thermometers to detect the temperature of critical components within the metering box or collect load current data using current sensors to determine if any abnormalities exist. These monitoring methods have been able to detect some equipment faults in practical applications, such as localized overheating caused by loose wiring terminals or abnormal fluctuations in load current, thus providing a certain level of assurance for the safe operation of the power system.

[0003] However, existing monitoring methods suffer from a technical deficiency in their insufficient analysis of the correlation between temperature anomalies and load conditions. In actual operation, terminal temperature anomalies are often the result of multiple factors, including poor contact, load fluctuations, and changes in ambient temperature. However, current technologies typically treat temperature monitoring and load analysis as independent monitoring methods, failing to establish a correlation between the two. This makes it difficult to accurately determine whether a temperature anomaly is due to equipment malfunction or normal temperature rise caused by load characteristics, thus affecting the accuracy of fault diagnosis and the timeliness of maintenance decisions. Consequently, maintenance personnel often need to conduct multiple on-site verifications and repeated tests when faced with temperature alarms, increasing maintenance costs and potentially delaying optimal fault handling, thus impacting power supply reliability. Summary of the Invention

[0004] The purpose of this invention is to at least partially solve one of the technical problems existing in the prior art.

[0005] To achieve the above objectives, the present invention provides a monitoring method for an energy metering box with temperature monitoring function, comprising:

[0006] Infrared thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box.

[0007] Load current analysis is performed on the power metering box based on three-phase current data to obtain the load characteristic curve, and the fluctuation characteristics of the load characteristic curve are extracted to obtain the load fluctuation parameters.

[0008] The operating status of the power metering box is evaluated based on the load fluctuation parameters to obtain the equipment health status, and the corresponding operation and maintenance strategy is executed based on the equipment health status.

[0009] Furthermore, infrared thermal imaging is performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map, including:

[0010] Thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain terminal thermal imaging data;

[0011] Based on the terminal thermal imaging data, the temperature values ​​of the incoming and outgoing terminals are calculated to obtain the terminal temperature matrix. Based on the terminal temperature matrix, the temperature gradient analysis of the terminal thermal imaging data is performed to obtain the terminal temperature distribution map.

[0012] Furthermore, based on the terminal temperature matrix, temperature gradient analysis is performed on the terminal thermal imaging data to obtain the terminal temperature distribution map, including:

[0013] Based on the terminal temperature matrix, the temperature difference between adjacent points of the terminal thermal imaging data is calculated to obtain a temperature gradient vector map. The gradient direction of the temperature gradient vector map is statistically analyzed to obtain a temperature change trend line.

[0014] The temperature gradient vector map is filtered by gradient amplitude threshold to obtain high gradient regions, and the center point of the region is located based on the high gradient regions to obtain the boundary of the heat accumulation area.

[0015] Based on the boundary of the heat accumulation area, the terminal thermal imaging data are overlaid and labeled to obtain a labeled temperature map. Then, based on the temperature change trend line, the labeled temperature map is overlaid with the trend direction to obtain the terminal temperature distribution map.

[0016] Furthermore, based on the terminal temperature matrix, temperature gradient analysis is performed on the terminal thermal imaging data to obtain the terminal temperature distribution map, including:

[0017] Based on the terminal temperature matrix, a horizontal differential operation is performed on each pixel in the terminal thermal imaging data to obtain a horizontal temperature difference component, and a vertical differential operation is performed on each pixel in the terminal thermal imaging data to obtain a vertical temperature difference component.

[0018] The gradient magnitude map is obtained by performing a summation and square root operation on the horizontal and vertical temperature difference components. The gradient direction angle map is obtained by calculating the arctangent angle on the horizontal and vertical temperature difference components.

[0019] The gradient direction angle map is divided into angle intervals and statistically analyzed to obtain the direction frequency distribution. Based on the direction frequency distribution, the peak direction is extracted to obtain the dominant gradient direction. Based on the dominant gradient direction, the high amplitude pixels in the gradient amplitude map are connected and fitted to obtain the temperature change trend line.

[0020] Furthermore, based on the temperature change trend line, the marked temperature map is superimposed with the trend direction to obtain the terminal temperature distribution map, including:

[0021] The endpoints of the temperature change trend line are extracted to obtain a trend endpoint set, and the direction angle is calculated based on the trend endpoint set to obtain the trend direction vector;

[0022] The labeled temperature map is aligned with the coordinate system based on the trend direction vector to obtain an alignment reference point, and the trend direction vector is projected into space based on the alignment reference point to obtain a trend projection layer.

[0023] The trend projection layer and the labeled temperature map are overlaid to obtain a fused temperature map, and the boundary contour of the fused temperature map is enhanced to obtain an enhanced boundary map.

[0024] Based on the enhanced boundary map, directional arrows are marked on the trend projection layer to obtain a directional annotation map, and the directional annotation map is rendered with color levels to obtain the terminal temperature distribution map.

[0025] Furthermore, based on the alignment reference point, the trend direction vector is spatially projected to obtain a trend projection layer, including:

[0026] Based on the alignment reference point, the image size of the labeled temperature map is read to obtain the image boundary range, and the scaling factor of the trend direction vector is calculated based on the image boundary range to obtain the projection scale factor.

[0027] Based on the projection scale factor, the starting point coordinates in the trend direction vector are scaled and translated to obtain the transformation starting point; and based on the projection scale factor, the ending point coordinates in the trend direction vector are scaled and translated to obtain the transformation ending point.

[0028] A connection path is generated based on the transformation start point and the transformation end point to obtain a vector line segment path, and a line width parameter is assigned to the vector line segment path to obtain a vector line segment diagram;

[0029] The vector line segment graph is configured with a transparency channel to obtain a transparent vector graphic, and the transparent vector graphic is encapsulated in a layer format to obtain the trend projection layer.

[0030] Furthermore, when the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box, including:

[0031] When the terminal temperature distribution map shows abnormal temperature spots, mark the coordinate position of the abnormal temperature spots in the terminal temperature distribution map;

[0032] Based on the coordinate position, a temperature anomaly trigger signal is generated and sent to the real-time monitoring system. Upon receiving the temperature anomaly trigger signal, the real-time monitoring system immediately starts collecting three-phase current data from the power metering box. The data is collected according to a preset sampling frequency and sampling duration to obtain the three-phase current data.

[0033] Furthermore, load current analysis is performed on the energy metering box based on three-phase current data to obtain load characteristic curves, including:

[0034] Based on the current of each phase in the three-phase current data, the effective value of the power metering box is calculated to obtain the current amplitude of each phase, and the current amplitude of each phase is vector synthesized to obtain the comprehensive load current value.

[0035] Based on the acquisition timestamps in the three-phase current data, the comprehensive load current value is time-series arranged to obtain current time-series data points, and the current time-series data points are checked at equal intervals to obtain a standard time-series current sequence.

[0036] The load characteristic curve is obtained by interpolating adjacent standard time-series current sequences.

[0037] Furthermore, fluctuation features are extracted from the load characteristic curve to obtain load fluctuation parameters, including:

[0038] Extreme point detection is performed on the load characteristic curve to obtain the peak point and valley point of the curve, and the difference between the peak point and valley point of the curve is calculated to obtain the current fluctuation amplitude.

[0039] The slope of the current fluctuation amplitude is calculated based on the peak point and valley point of the curve to obtain the current change rate, and the current change rate is statistically calculated to obtain the mean rate and the peak rate.

[0040] The load fluctuation parameters are obtained by integrating the current fluctuation amplitude, the average rate, and the peak rate.

[0041] Furthermore, the operating status of the power metering box is evaluated based on the load fluctuation parameters to obtain the equipment health status, including:

[0042] The load fluctuation parameters are divided into thresholds to obtain fluctuation level values, and the fluctuation level values ​​are statistically analyzed over time to obtain the fluctuation duration.

[0043] The fluctuation level value is weighted based on the duration of the fluctuation to obtain a load state index, and the power metering box is described based on the load state index to obtain the equipment health status.

[0044] The present invention also provides an energy metering box with temperature monitoring function, comprising:

[0045] The scanning module is used to perform infrared thermal imaging scanning on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box.

[0046] The analysis module is used to perform load current analysis on the power metering box based on three-phase current data, obtain the load characteristic curve, and extract the fluctuation characteristics of the load characteristic curve to obtain the load fluctuation parameters.

[0047] The evaluation module is used to evaluate the operating status of the power metering box based on the load fluctuation parameters, obtain the equipment health status, and execute the corresponding operation and maintenance strategy based on the equipment health status.

[0048] This invention provides a monitoring method for an energy metering box with temperature monitoring function, comprising: performing infrared thermal imaging scanning on the incoming and outgoing terminals of the energy metering box to obtain a terminal temperature distribution map; when the terminal temperature distribution map shows abnormal temperature spots, triggering a real-time monitoring system to collect three-phase current data of the energy metering box; performing load current analysis on the energy metering box based on the three-phase current data to obtain a load characteristic curve, and extracting fluctuation characteristics from the load characteristic curve to obtain load fluctuation parameters; evaluating the operating status of the energy metering box based on the load fluctuation parameters to obtain the equipment health status, and based on the equipment health status... Kang State executes corresponding operation and maintenance strategies, solving the technical problem that existing technologies typically treat temperature monitoring and load analysis as independent monitoring methods, failing to establish a correlation between the two. This makes it difficult to accurately determine whether the temperature rise is caused by equipment failure or normal temperature rise due to load characteristics when an abnormal temperature is detected, thus affecting the accuracy of fault diagnosis and the timeliness of operation and maintenance decisions. By performing load characteristic analysis and fluctuation parameter extraction on three-phase current data, it can identify the correlation between load change patterns and temperature anomalies, accurately distinguishing whether the temperature rise is caused by equipment failure or normal heating due to load characteristics, reducing false alarm rates and improving the accuracy of fault diagnosis. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 This is a schematic diagram of the steps of a monitoring method for an energy metering box with temperature monitoring function in one embodiment of the present invention;

[0051] Figure 2 This is a schematic diagram of an energy metering box with temperature monitoring function in one embodiment of the present invention;

[0052] The objectives, features, and advantages of this invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0053] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention. The step numbers in the following embodiments are set only for ease of explanation, and there is no limitation on the order between the steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0054] The following describes in detail, with reference to the accompanying drawings, a method for monitoring an energy metering box with temperature monitoring function according to an embodiment of the present invention. First, the method for monitoring an energy metering box with temperature monitoring function according to an embodiment of the present invention will be described in detail with reference to the accompanying drawings.

[0055] Figure 1 This invention provides a method for monitoring an energy metering box with temperature monitoring function, comprising:

[0056] Step S1: Perform infrared thermal imaging scanning on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, trigger the real-time monitoring system to collect the three-phase current data of the power metering box.

[0057] Specifically, when monitoring the electricity metering box, an infrared thermal imager is first used to scan the incoming and outgoing terminal areas. During this process, the imager's detector receives the infrared energy radiated from the terminal surface. The system converts these infrared radiation signals into electrical signals to generate a terminal temperature distribution map. Different colors in the map represent different temperature ranges; for example, red areas have higher temperatures, while blue areas have lower temperatures. The monitoring system analyzes the generated temperature distribution map in real time, judging whether there are abnormal temperature spots by using a set temperature threshold. This threshold is usually determined based on the temperature range of the terminal during normal operation. For example, when the temperature of a certain terminal area is more than 15°C higher than the surrounding ambient temperature, that area will form a noticeable high-temperature spot in the temperature distribution map. Once such an abnormal temperature spot is detected, the system immediately sends a trigger signal to the real-time monitoring module. After receiving the signal, the monitoring module activates the current acquisition device, which synchronously acquires the current data of phases A, B, and C through the current transformer installed in the metering box. The acquisition frequency is usually set to more than 10 times per second. This completes the linkage process from temperature anomaly detection to current data acquisition, with the entire trigger response time controlled within 2 seconds, ensuring that the load status at the moment of an anomaly is captured.

[0058] Step S2: Based on the three-phase current data, perform load current analysis on the power metering box to obtain the load characteristic curve, and extract the fluctuation characteristics of the load characteristic curve to obtain the load fluctuation parameters.

[0059] Specifically, after the system collects current data for phases A, B, and C, it arranges this data according to a time series. For example, within a 5-minute monitoring window, collecting data 10 times per second yields 3000 data points. These current values ​​are then plotted as a curve over time, with time on the horizontal axis and current amplitude on the vertical axis, forming the load characteristic curve. Next, fluctuation characteristics are extracted from the curve. First, the current difference between adjacent data points on the curve is calculated. For example, if the current in phase A is 50 amps at one moment and changes to 55 amps the next second, the difference is 5 amps. All these differences are recorded, and their standard deviation is calculated. The larger the standard deviation, the more severe the current fluctuation. Additionally, the frequency of peak occurrences on the curve is counted, specifically how many times peaks exceeding 20% ​​of the average value appeared during the monitoring period. This frequency reflects the frequency of load impacts. Finally, the rate of change of current must be calculated. The current value at the current moment is subtracted from the current value at the previous moment and then divided by the time interval. The resulting value represents the speed at which the current rises or falls. These three parameters combined are the load fluctuation parameters, which can comprehensively describe the fluctuation of the load current and provide data basis for subsequent judgment of the cause of abnormal temperature.

[0060] Step S3: Evaluate the operating status of the power metering box based on the load fluctuation parameters to obtain the equipment health status, and execute the corresponding operation and maintenance strategy based on the equipment health status.

[0061] Specifically, after obtaining the load fluctuation parameters, the system compares these parameters with preset judgment standards to evaluate the operating status. Specifically, a scoring table is established. If the standard deviation is less than 3 amps, the peak frequency does not exceed 2 times per minute, and the current change rate is less than 8 amps per second, it is judged to be in a healthy state and scores above 80 points. If the standard deviation is between 3 and 6 amps or the peak frequency reaches 3 to 5 times per minute, it is in a sub-healthy state and scores between 60 and 80 points. If the standard deviation exceeds 6 amps and the peak frequency exceeds 5 times per minute, the equipment health status is defined as a fault state and scores below 60 points. After the assessment is completed, the system will automatically match the operation and maintenance strategy according to the obtained equipment health status. For example, in a healthy state, it is only necessary to maintain the regular monthly inspection cycle. In a sub-healthy state, the inspection interval should be shortened to once a week and the temperature monitoring frequency should be increased. In a fault state, a work order should be immediately dispatched to notify the operation and maintenance personnel to carry out on-site repairs. At the same time, the system will push the terminal temperature distribution map and load characteristic curve to the maintenance personnel to facilitate them to quickly locate the problem terminal. In this way, the entire process from parameter analysis to strategy execution is connected.

[0062] In a specific embodiment, infrared thermal imaging scanning is performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map, including:

[0063] Thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain terminal thermal imaging data;

[0064] Based on the terminal thermal imaging data, the temperature values ​​of the incoming and outgoing terminals are calculated to obtain the terminal temperature matrix. Based on the terminal temperature matrix, the temperature gradient analysis of the terminal thermal imaging data is performed to obtain the terminal temperature distribution map.

[0065] Specifically, in actual operation, the process of obtaining the terminal temperature distribution map needs to be completed in several steps. First, an infrared thermal imager is used to scan the incoming and outgoing terminal areas of the power metering box. These incoming and outgoing terminals include the A-phase, B-phase, and C-phase incoming terminals connected to the power supply side and the three-phase outgoing terminals output from the load side. The focal plane detector of the thermal imager will receive the infrared energy radiated from the surface of these terminals point by point. The array of thermal elements in the detector is usually 320×240 or 640×480 pixels, with each pixel corresponding to a small area on the terminal surface. During scanning, the thermal imager records the received infrared radiation intensity to form raw data. This data is the terminal thermal imaging data, which is essentially a digital matrix containing infrared radiation intensity values.

[0066] After obtaining the terminal thermal imaging data, temperature calculations are required. Since thermal imaging data records radiation intensity rather than temperature values, Planck's law of radiation is used for conversion. The system reads the radiation intensity value of each pixel and then corrects it according to the pre-set emissivity parameters of the terminal material. The emissivity of copper terminals is generally around 0.78. The calculation formula also needs to consider the influence of ambient temperature and atmospheric transmittance. For example, if the ambient temperature is 25℃ and the atmospheric transmittance is calculated as 0.95, the radiation intensity of each pixel can be converted into the corresponding temperature value. Suppose the radiation intensity of a certain pixel is calculated to be 62℃, then this value is filled into the corresponding position in the terminal temperature matrix. After the entire terminal area is scanned, a temperature matrix with the same specifications as the pixel array of the thermal imager is obtained. Each element of the matrix stores the specific temperature value.

[0067] Having the terminal temperature matrix isn't enough to directly generate a distribution map; a temperature gradient analysis is needed first to highlight abnormal temperature points. This analysis calculates the temperature difference between each location in the matrix and its surroundings. For example, if a given element has a temperature of 58℃, and its four adjacent elements have temperatures of 56℃, 57℃, 55℃, and 56℃ respectively, the temperature gradient at that location is calculated by subtracting the average temperature of the surrounding area (56.5℃) from 58℃, resulting in a gradient of 1.5℃. However, if a terminal has poor contact causing localized overheating, its temperature might reach 75℃ while the surrounding area is only around 58℃. In this case, the calculated temperature gradient would be as large as 17℃. After calculating the temperature gradients for all locations, the system assigns color codes to different areas based on the gradient magnitude. Areas with a temperature gradient less than 2℃ are displayed in blue or green, indicating uniform temperature distribution; gradients between 2 and 5℃ are displayed in yellow; and gradients exceeding 5℃ are marked in orange or red. The resulting color image after processing the terminal thermal imaging data is the terminal temperature distribution map. The map clearly shows which locations have abnormal temperature spots, allowing maintenance personnel to easily identify problem areas.

[0068] In a specific embodiment, temperature gradient analysis is performed on the terminal thermal imaging data based on the terminal temperature matrix to obtain the terminal temperature distribution map, including:

[0069] Based on the terminal temperature matrix, the temperature difference between adjacent points of the terminal thermal imaging data is calculated to obtain a temperature gradient vector map. The gradient direction of the temperature gradient vector map is statistically analyzed to obtain a temperature change trend line.

[0070] The temperature gradient vector map is filtered by gradient amplitude threshold to obtain high gradient regions, and the center point of the region is located based on the high gradient regions to obtain the boundary of the heat accumulation area.

[0071] Based on the boundary of the heat accumulation area, the terminal thermal imaging data are overlaid and labeled to obtain a labeled temperature map. Then, based on the temperature change trend line, the labeled temperature map is overlaid with the trend direction to obtain the terminal temperature distribution map.

[0072] Specifically, when performing temperature gradient analysis based on the terminal temperature matrix, the first step is to calculate the temperature difference between adjacent points. This calculation is performed for each position in the matrix. Taking the point at coordinate (i, j) in the matrix as an example, its temperature is denoted as T(i, j). The system will read the temperature values ​​of its four neighboring points (up, down, left, and right) as T(i-1, j), T(i+1, j), T(i, j-1), and T(i, j+1), respectively. Then, the lateral temperature difference ΔTx is calculated as T(i, j+1) minus T(i, j-1) divided by 2, and the longitudinal temperature difference ΔTx is calculated as T(i, j+1) minus T(i, j-1) divided by 2. Ty equals T(i+1,j) minus T(i-1,j) divided by 2. For example, if the temperature at a certain terminal is 68℃, its right neighbor is 72℃, and its left neighbor is 65℃, then the lateral temperature difference is (72-65) / 2, which is 3.5℃. After obtaining the longitudinal temperature difference using the same algorithm, these two temperature difference values ​​are combined into a vector. The lateral temperature difference is the x-component of the vector, and the longitudinal temperature difference is the y-component. After processing each point in the matrix in this way, a temperature gradient vector diagram is formed. Each position on the diagram has an arrow indicating the direction of the fastest temperature increase.

[0073] After obtaining the temperature gradient vector map, gradient direction statistics are also performed. The system will iterate through all the arrows in the vector map, calculate the angle between each arrow and the horizontal direction, with the angle range from 0 to 360 degrees, and then divide these angles into eight groups of 45-degree intervals, corresponding to the eight directions of east, southeast, south, southwest, west, northwest, north, and northeast. After the statistics are completed, if it is found that there are a lot of arrows in a certain direction, such as 60% of the arrows pointing to the southeast, it means that the temperature of the entire terminal area is increasing in the southeast direction. The system will draw a temperature change trend line on the vector map. The direction of this line is determined according to the dominant direction found in the statistics. It reflects the diffusion path of heat on the terminal surface. If a certain A-phase incoming terminal heats up at the connection bolt position due to poor contact, the temperature will diffuse from the center of the bolt to the surrounding area, and the trend line will show a radial distribution.

[0074] The next step is to filter out high gradient regions from the temperature gradient vector map. This step requires setting a gradient magnitude threshold. The gradient magnitude is calculated by taking the square root of the sum of the squares of the previously obtained lateral and longitudinal temperature differences. For example, if the lateral temperature difference at a certain location is 3.5℃ and the longitudinal temperature difference is 2.8℃, then the gradient magnitude is approximately 4.5℃ / pixel (3.5² + 2.8²). The system sets the threshold at 3℃ / pixel. Any location where the gradient magnitude exceeds this threshold is marked as a high gradient region. These regions often correspond to areas where the terminal surface temperature changes drastically, which are potential locations of faults. After filtering, the system will locate the center point of each high gradient region. The method is to find the pixel with the highest temperature in the region as the center, and then search outwards from the center point. When the temperature in a certain direction is more than 5°C lower than the center point, the search stops. Connecting the stop points in all directions gives the boundary of the heat accumulation area. This boundary delineates the range of abnormal heating. For example, if the bolt of a certain outgoing terminal is loose, the contact resistance will increase and the temperature of the heating center will reach 82°C. The temperature at the boundary will drop to about 76°C. The system can then draw an elliptical boundary to frame this heat accumulation area.

[0075] After determining the boundary of the heat accumulation area, it needs to be superimposed on the terminal thermal imaging data. The system will draw the boundary with red or orange lines on the original thermal image, and fill the boundary with semi-transparent warning blocks. This process produces the labeled temperature map, which retains the original temperature distribution information and highlights abnormal areas. However, this is not the final terminal temperature distribution map. The previously calculated temperature change trend line also needs to be superimposed. The system will draw the trend line on the labeled temperature map with arrows or streamlines. The direction of the arrows indicates the direction of heat diffusion. This way, maintenance personnel can see not only where the temperature is high, but also the direction from which the heat is coming. For example, if a trend line is found next to a high-temperature area of ​​a terminal pointing to an adjacent terminal, it indicates that the heat may be caused by the adjacent terminal rather than a fault in the terminal itself. After the trend direction is superimposed, the output image is the terminal temperature distribution map.

[0076] In a specific embodiment, temperature gradient analysis is performed on the terminal thermal imaging data based on the terminal temperature matrix to obtain the terminal temperature distribution map, including:

[0077] Based on the terminal temperature matrix, a horizontal differential operation is performed on each pixel in the terminal thermal imaging data to obtain a horizontal temperature difference component, and a vertical differential operation is performed on each pixel in the terminal thermal imaging data to obtain a vertical temperature difference component.

[0078] The gradient magnitude map is obtained by performing a summation and square root operation on the horizontal and vertical temperature difference components. The gradient direction angle map is obtained by calculating the arctangent angle on the horizontal and vertical temperature difference components.

[0079] The gradient direction angle map is divided into angle intervals and statistically analyzed to obtain the direction frequency distribution. Based on the direction frequency distribution, the peak direction is extracted to obtain the dominant gradient direction. Based on the dominant gradient direction, the high amplitude pixels in the gradient amplitude map are connected and fitted to obtain the temperature change trend line.

[0080] Specifically, during the temperature gradient analysis of the terminal temperature matrix, the system performs a differential operation on each pixel in the terminal thermal imaging data. The horizontal differential operation involves subtracting the temperature value of the left neighbor from the temperature value of the right neighbor of a pixel. For example, if the coordinates of a position in the terminal temperature matrix are (m, n), its horizontal temperature difference component is calculated by subtracting T(m, n+1) from T(m, n-1). In a practical example, if the temperature to the right of the pixel corresponding to the bolt position of a phase A incoming terminal is 7... If the temperature to the left of 3℃ is 67℃, then the horizontal temperature difference component at this point is 6℃. After processing each pixel in the matrix in this way, we get a set of horizontal temperature difference component data. Similarly, the vertical difference operation is to subtract the temperature of the upper neighbor from the temperature of the lower neighbor of the pixel. In terms of formula, it is T(m+1, n) minus T(m-1, n). For example, if the temperature below a certain position is 70℃ and the temperature above is 64℃, the vertical temperature difference component is calculated to be 6℃. These two sets of temperature difference components reflect the rate of temperature change in the horizontal and vertical directions, respectively.

[0081] After obtaining the horizontal and vertical temperature difference components, the gradient magnitude needs to be calculated. The calculation method is to square the two components, add them together, and then take the square root. Continuing with the previous example, substituting the horizontal temperature difference component of 6℃ and the vertical temperature difference component of 6℃ into the formula, we get √(6²+6²), which calculates to approximately 8.5℃. This value is the gradient magnitude of that pixel, representing the degree of temperature change at that location. After processing the entire terminal temperature matrix, a gradient magnitude map is formed. The larger the value on the map, the steeper the temperature change. In addition, the gradient direction angle also needs to be calculated using the arctangent function. The vertical temperature difference component is divided by the horizontal temperature difference component, and the arctangent value is taken. In the previous example, 6 divided by 6 equals 1, and the arctangent of 1 gives 45 degrees. This is the gradient direction angle of that point, indicating the direction of the fastest temperature rise. After calculating all pixels in this way, a gradient direction angle map is generated. Each position on the map has an angle value between 0 and 360 degrees.

[0082] Next, the gradient orientation angle map needs to be divided into angle intervals for statistical analysis. The system divides the 360-degree range into eight intervals, each spanning 45 degrees, corresponding to 0 to 45 degrees, 45 to 90 degrees, 90 to 135 degrees, and so on up to 315 to 360 degrees. Then, the gradient orientation angle map is traversed to count the number of pixels appearing in each interval. This statistical result is the orientation frequency distribution. For example, after the statistics are completed, it is found that there are 820 pixels in the 0 to 45 degree interval, 350 pixels in the 45 to 90 degree interval, and 180 pixels in the 90 to 135 degree interval. With fewer pixels in other ranges, the 0 to 45 degree range has the highest frequency of occurrence. The system will identify the range with the most occurrences from the directional frequency distribution and define the middle angle of this range as the dominant gradient direction. For example, the middle angle of the 0 to 45 degree range is 22.5 degrees. This angle represents the main direction of temperature change in the entire terminal area. If a terminal is heated due to a short circuit current impact on the line side, the heat may spread along the direction of the conductor, and the dominant gradient direction will be basically consistent with the direction of conductor arrangement.

[0083] After determining the dominant gradient direction, high-amplitude pixels need to be identified from the gradient magnitude map. The system sets a threshold, such as 7°C, and filters out all pixels in the gradient magnitude map with values ​​exceeding 7°C. These points are usually distributed in regions of drastic temperature changes. Then, these high-amplitude pixels are connected to form a line to fit the temperature change trend. The rule for connecting the lines is to prioritize connecting pixels whose gradient direction angle is close to the dominant gradient direction. Specifically, the system starts from the pixel with the largest gradient magnitude and finds eight neighboring pixels whose gradient direction angle differs from 22.5 degrees by no more than 30 degrees. These neighboring pixels that meet the criteria are then connected. Connect the points, and then continue searching outwards from this neighboring point to find the next matching point. This process continues until a curve is formed. For example, if the wiring bolt of a certain B-phase incoming terminal is loose, and the gradient amplitude at the center of the bolt reaches 12°C with a gradient direction angle of 25 degrees, the system starts from this center point and searches northeastwards. It finds a nearby point with a gradient amplitude of 9.5°C and a direction angle of 18 degrees and connects to it. Then it continues searching and connecting to a point with a gradient amplitude of 8.2°C and a direction angle of 32 degrees. The resulting curve shows that the heat is spreading outwards from the center of the bolt along the northeast direction. This fitted curve is the temperature change trend line.

[0084] In a specific embodiment, the terminal temperature distribution map is obtained by superimposing the labeled temperature map based on the temperature change trend line, including:

[0085] The endpoints of the temperature change trend line are extracted to obtain a trend endpoint set, and the direction angle is calculated based on the trend endpoint set to obtain the trend direction vector;

[0086] The labeled temperature map is aligned with the coordinate system based on the trend direction vector to obtain an alignment reference point, and the trend direction vector is projected into space based on the alignment reference point to obtain a trend projection layer.

[0087] The trend projection layer and the labeled temperature map are overlaid to obtain a fused temperature map, and the boundary contour of the fused temperature map is enhanced to obtain an enhanced boundary map.

[0088] Based on the enhanced boundary map, directional arrows are marked on the trend projection layer to obtain a directional annotation map, and the directional annotation map is rendered with color levels to obtain the terminal temperature distribution map.

[0089] Specifically, to overlay the temperature trend line onto the labeled temperature map to generate the final terminal temperature distribution map, the endpoints of the temperature trend line must first be extracted. Since the trend line is composed of multiple connected segments, the system reads the starting and ending coordinates of each segment one by one. For example, if a trend line extends from the connection bolt position of the A-phase incoming terminal, the first segment's starting coordinates are (156, 203) and its ending coordinates are (178, 215). The second segment's starting coordinates continue from the previous segment's ending coordinates, from (178, 215) to (195, 228). After extracting the endpoint coordinates of all segments, they are placed in an array; this array is the trend endpoint set. Next, we need to calculate the direction angle of each line segment. The calculation method is to subtract the y-coordinate of the starting point from the y-coordinate of the end point of the line segment to obtain the vertical difference, and then subtract the x-coordinate of the starting point from the x-coordinate of the end point to obtain the horizontal difference. Taking the line segment from (156, 203) to (178, 215) as an example, the vertical difference is 215-203, which equals 12, and the horizontal difference is 178-156, which equals 22. The angle calculated using the arctangent function is arctan(12 / 22), which is approximately 28.6 degrees. After calculating each line segment in this way, we get a set of angle values. These angle values ​​are combined with the corresponding line segment lengths to form a trend direction vector. The magnitude of the vector is the line segment length, and the direction is the angle just calculated.

[0090] Next, the trend direction vector needs to be aligned with the labeled temperature map. Since the trend line is extracted from the gradient magnitude map, while the labeled temperature map is obtained by marking the boundaries of heat accumulation areas based on terminal thermal imaging data, the coordinate systems of the two maps may be misaligned. The alignment operation requires finding a reference point. The system will find the pixel with the highest temperature on the labeled temperature map as the alignment reference point. Let's assume the coordinates of this point are (168, 210). Simultaneously, it will find the endpoint in the trend endpoint set closest to the location of the maximum gradient magnitude, for example, with coordinates (165, 207). Then, the alignment reference point and the trend endpoint... The coordinate offset is 3 pixels in the x-direction and 3 pixels in the y-direction. Then, this offset is added to the coordinates of all trend direction vectors to complete the alignment. After alignment, spatial position projection is performed. The system will create a transparent layer with the same size as the labeled temperature map. The adjusted trend direction vectors are projected onto this transparent layer according to the coordinates of the vector's starting point. During projection, each vector will be drawn as a directional line segment. The color of the line segment is determined by the magnitude of the vector. The larger the magnitude, the darker the color. This transparent layer is the trend projection layer, which separately stores the directional information of the temperature change trend.

[0091] After generating the trend projection layer, it needs to be overlaid with the labeled temperature map. The labeled temperature map already outlines the boundaries of the heat accumulation area with red lines. Now, the trend projection layer is overlaid on top. The overlay of the two layers uses an opacity blending algorithm, with the opacity of the labeled temperature map set to 80% and the opacity of the trend projection layer set to 50%. The resulting fused temperature map shows both the original temperature distribution and boundary labels, as well as the overlaid trend lines. However, the boundary outlines may become less clear due to the opacity blending, so boundary outline enhancement processing is required. The system will detect the boundary lines of the heat accumulation area in the fused temperature map, use the Sobel operator to perform convolution operations on the pixels near the boundary to find the edge positions, and then increase the color saturation of the boundary lines by 30% and thicken the line width from the original 2 pixels to 3 pixels. After processing, the output is the enhanced boundary map, and the boundaries of the heat accumulation area on the map become more prominent.

[0092] After the boundary map is generated, directional arrows need to be labeled on the line segments of the trend projection layer. Line segments alone are not intuitive enough; adding arrows allows maintenance personnel to immediately see the direction of heat dissipation. The system adds a triangular arrow to the end of each line segment in the trend projection layer. The direction of the arrow indicates the direction of the line segment. The arrow size is set proportionally to the line segment length. For example, if a line segment is 35 pixels long, the corresponding arrow's base width is set to 7 pixels and its height to 10 pixels. For a real-world example, suppose the connection bolts of a phase B output terminal have oxidized contact surfaces due to long-term operation. The heat spreads along the northeast direction from the heating center at 82℃. The system will draw an elliptical red border around the 82℃ high-temperature area on the enhanced boundary map, and simultaneously overlay a trend line pointing northeast inside the border with an arrow at the end of the line segment. This process produces a directional annotation map. The final step is color gradation rendering output. The system renders the directional annotation map according to a preset color gradation scheme: areas with temperatures below 50℃ are displayed in blue, 50 to 65℃ in green, 65 to 75℃ in yellow, and above 75℃ in red. The output image after rendering is the terminal temperature distribution map.

[0093] In a specific embodiment, a trend projection layer is obtained by spatially projecting the trend direction vector based on the alignment reference point, including:

[0094] Based on the alignment reference point, the image size of the labeled temperature map is read to obtain the image boundary range, and the scaling factor of the trend direction vector is calculated based on the image boundary range to obtain the projection scale factor.

[0095] Based on the projection scale factor, the starting point coordinates in the trend direction vector are scaled and translated to obtain the transformation starting point; and based on the projection scale factor, the ending point coordinates in the trend direction vector are scaled and translated to obtain the transformation ending point.

[0096] A connection path is generated based on the transformation start point and the transformation end point to obtain a vector line segment path, and a line width parameter is assigned to the vector line segment path to obtain a vector line segment diagram;

[0097] The vector line segment graph is configured with a transparency channel to obtain a transparent vector graphic, and the transparent vector graphic is encapsulated in a layer format to obtain the trend projection layer.

[0098] Specifically, when projecting the trend direction vector into space, the system first uses the alignment reference point to read the image size of the labeled temperature map. This reading operation calls the interface function of the image processing library to obtain the width and height pixel values ​​of the image. If the size of the labeled temperature map is 640 pixels wide and 480 pixels high, then the image boundary range is a rectangular area from 0 to 640 on the x-axis and from 0 to 480 on the y-axis. After obtaining the image boundary range, the projection scale factor needs to be calculated. Because the trend direction vector is originally in the coordinate system of the gradient magnitude map, the size of the gradient magnitude map may be different from that of the labeled temperature map. For example, if the gradient magnitude map is 320×240, then the scale factor in the x-direction is calculated by dividing 640 by 320, which is 2.0, and the scale factor in the y-direction is calculated by dividing 480 by 240, which is also 2.0. The combination of these two values ​​is the projection scale factor, which indicates that one pixel on the gradient magnitude map corresponds to two pixels on the labeled temperature map.

[0099] Next, we need to use the projection scale factor to scale and translate the coordinates in the trend direction vector. Each trend direction vector has a start and end coordinates. Taking the start coordinates (85, 103) of a certain vector as an example, the scaling transformation is to multiply the x coordinate by 2.0 to get 170 and the y coordinate by 2.0 to get 206. Then we need to add the previously calculated alignment offset. As mentioned earlier, the offset of the alignment reference point is 3 pixels in the x direction and 3 pixels in the y direction. So the final coordinates of the transformed start point are (170+3, 206+3), which equals (173, 209). The same method is used to process the end coordinates. If the original end point is (97, 115), after scaling it is (194, 230). Adding the offset, we get the transformed end point (197, 233). All the start and end points in the entire trend endpoint set are processed in this way, and the mapping from the gradient magnitude map coordinate system to the labeled temperature map coordinate system is completed.

[0100] After transforming the start and end points, a connecting path needs to be generated. The system will traverse all transformation start and end points, connecting the corresponding start and end points with straight lines. The Bressenham line algorithm is used to determine the pixels the line segment passes through. For example, to draw a straight line from the transformation start point (173, 209) to the transformation end point (197, 233), the algorithm will calculate that the line segment passes through 24 pixels: (173, 209), (174, 210), (175, 211), and so on, up to (197, 233). Storing the coordinates of all these pixels in order forms the vector line path. This path records the complete direction of the line segment on the image. Then, the line width parameter needs to be set for the vector line segment path. The line width is assigned according to the magnitude of the vector. For example, if the magnitude of the vector corresponding to a line segment is 18 pixels, then the line width is set to 2 pixels. If the magnitude of the vector is 35 pixels, then the line width is set to 3 pixels. The system expands each pixel on the path by half the line width in the direction perpendicular to the line segment. For example, if the line width is 2 pixels, then it is expanded by 1 pixel on each side. After this processing, the path that was originally one pixel wide becomes a line segment with actual width. After the line width is assigned, the output is the vector line segment image, on which you can see line segments of different thicknesses.

[0101] After generating the vector line segment map, a transparency channel needs to be configured because it will be overlaid on the temperature map later. The transparency channel makes the areas outside the line segments completely transparent. The configuration involves adding an alpha channel to the vector line segment map, which exists alongside the RGB color channels. The alpha value of the line segment pixels is set to 255 to indicate complete opacity, while the alpha value of the background pixels outside the line segments is set to 0 to indicate complete transparency. This configuration results in a transparent vector graphic, where only the line segments are visible; the rest is transparent. Finally, the transparent vector graphic needs to be encapsulated into a layer format. The encapsulation process organizes the data according to the layer data structure of the image processing software, including the layer's width, height, pixel data, and transparency channel. The data and layer blending modes, etc., are usually set to "Normal" to indicate direct overlay. After encapsulation, the output is the trend projection layer, which can be easily overlaid with other layers as an independent layer. This trend projection layer is a refined implementation of the technical element of "obtaining the trend projection layer" in the upper-level solution. Through a series of steps such as size reading, scale calculation, coordinate transformation, path generation, line width assignment, transparency configuration, and format encapsulation, the conversion process from trend direction vector to overlayable layer is completed. If the wiring bolt of a C-phase incoming terminal is loose, causing local heating, the trend projection layer will display a line segment radiating outward from the center of the bolt at the corresponding position. The thickness of the line segment reflects the intensity of temperature diffusion and the direction indicates the heat propagation path.

[0102] In a specific embodiment, when the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box, including:

[0103] When the terminal temperature distribution map shows abnormal temperature spots, mark the coordinate position of the abnormal temperature spots in the terminal temperature distribution map;

[0104] Based on the coordinate position, a temperature anomaly trigger signal is generated and sent to the real-time monitoring system. Upon receiving the temperature anomaly trigger signal, the real-time monitoring system immediately starts collecting three-phase current data from the power metering box. The data is collected according to a preset sampling frequency and sampling duration to obtain the three-phase current data.

[0105] Specifically, after generating the terminal temperature distribution map, the system performs real-time image detection to determine if there are any abnormal temperature spots. The detection method involves traversing all pixels of the terminal temperature distribution map and reading the color value of each pixel. Because the temperature and color have already been correlated during the color rendering process, areas with temperatures above 75℃ are displayed as red or orange. The system identifies pixels with a red component greater than 200 and a green component less than 100 in their RGB values ​​as high-temperature pixels. When a continuous area of ​​high-temperature pixels is found to exceed 50 pixels, it is determined that there are abnormal temperature spots. For a practical example, suppose a certain A-phase incoming terminal... A poor connection in the wiring bolt creates a red spot on the terminal temperature distribution map. Upon detecting this spot, the system immediately marks its coordinates. This marking process involves calculating the centroid coordinates of the spot area. Specifically, the x-coordinates of all pixels within the spot are summed and divided by the total number of pixels to obtain the centroid's x-coordinate. The y-coordinate is calculated using the same method. For example, if a spot contains 86 pixels, the sum of its x-coordinates is 15308, divided by 86 to obtain the centroid's x-coordinate of 178, and the sum of its y-coordinates is 18103, divided by 86 to obtain the centroid's y-coordinate of 210. Therefore, the coordinates of this abnormal temperature spot on the terminal temperature distribution map are marked as (178, 210).

[0106] After marking the coordinates, the system generates a temperature anomaly trigger signal based on these coordinates. The trigger signal is a data packet containing information such as an anomaly type identifier, anomaly point coordinates, detection timestamp, and temperature value. The anomaly type identifier is encoded using a one-byte value; for example, 0x01 indicates a single-point high-temperature anomaly, and 0x02 indicates a multi-point high-temperature anomaly. The anomaly point coordinates are the previously marked (178, 210). The detection timestamp records the exact moment the trigger signal was generated, accurate to milliseconds. The temperature value is read from the corresponding position on the terminal temperature distribution map. For example, if the temperature at coordinates (178, 210) is 82℃, this value is written into the data packet. After the data packet is assembled, it is sent to the real-time monitoring system through a communication interface. The communication interface can be an RS485 bus, Ethernet, or a wireless module. The transmission uses an agreed-upon communication protocol. The protocol frame includes a start flag, data length, data content, and checksum. The start flag uses a fixed byte sequence 0xAA. 0x55 indicates that the data length field records the number of bytes of subsequent data. The data content is the previously assembled trigger signal data packet. The checksum is calculated using the CRC16 algorithm on the entire data frame, so that the temperature anomaly trigger signal sent out can be correctly received and parsed by the real-time monitoring system.

[0107] After receiving a temperature anomaly trigger signal, the real-time monitoring system first performs a checksum verification to confirm that the data transmission process is error-free. Then, it parses the data packet to extract the coordinates of the anomaly point and the temperature value. Next, it immediately starts the three-phase current data acquisition program. The start operation involves sending a start command to the current acquisition module. Upon receiving the command, the current acquisition module opens the signal channels of the three current transformers (A, B, and C phases). The secondary side output current signal of the current transformer is converted into a 0- to 5-volt voltage signal by a signal conditioning circuit and then sent to an analog-to-digital converter for digitization. Data acquisition must be performed according to a preset sampling frequency and sampling duration. The sampling frequency is typically set to 10 times per second, or one point every 100 milliseconds. This frequency is sufficient to capture... For short-term fluctuations in load current, a sampling duration of 5 minutes means continuous sampling for 300 seconds starting from the moment the trigger signal is received. During these 5 minutes, 10 points are sampled per second for each current channel, resulting in a total of 3000 data points. For example, when the temperature of the B-phase outgoing terminal becomes abnormal due to a short circuit impact on the load side, the system receives the trigger signal and immediately starts sampling. In the following 5 minutes, it continuously records the changes in the B-phase current, which may capture the process of the current suddenly jumping from the normal 50 amps to 180 amps and then falling back. After the sampling is completed, the data from the A-phase, B-phase, and C-phase channels are stored in three arrays, each containing 3000 current values. These three arrays are combined to obtain the three-phase current data.

[0108] In a specific embodiment, load current analysis is performed on the power metering box based on three-phase current data to obtain a load characteristic curve, including:

[0109] Based on the current of each phase in the three-phase current data, the effective value of the power metering box is calculated to obtain the current amplitude of each phase, and the current amplitude of each phase is vector synthesized to obtain the comprehensive load current value.

[0110] Based on the acquisition timestamps in the three-phase current data, the comprehensive load current value is time-series arranged to obtain current time-series data points, and the current time-series data points are checked at equal intervals to obtain a standard time-series current sequence.

[0111] The load characteristic curve is obtained by interpolating adjacent standard time-series current sequences.

[0112] Specifically, after obtaining the three-phase current data, load current analysis needs to be performed on the electricity metering box. First, the effective value of each phase current must be calculated. Because the collected raw current data are instantaneous values, with one point sampled every 100 milliseconds, these instantaneous values ​​fluctuate with the periodic changes in AC power. Therefore, it is necessary to calculate the effective value that represents the current magnitude at that moment. The calculation method is to take the squares of the sampling points within one power frequency cycle, sum them, and then take the square root. my country's power grid frequency is 50 Hz, and one cycle is 20 milliseconds. At a sampling frequency of 10 times per second, only 2 points can be collected within one cycle, which is a low sampling rate. In actual operation, the system will trigger... During data acquisition, the sampling frequency is temporarily increased to 1000 times per second, or one point per millisecond. This results in 20 sampling points within a power frequency cycle. If the 20 instantaneous values ​​of phase A current in a certain cycle are 48.2, 51.6, 54.3 and up to 47.8 amperes, the squares of these 20 values ​​are summed to get 51236. Dividing by 20 gives 2561.8. Taking the square root gives the effective value of 50.6 amperes. Phases B and C are processed in the same way. For example, the effective value of phase B is calculated to be 49.8 amperes and phase C to be 51.2 amperes. These three values ​​are the current amplitudes of each phase.

[0113] After obtaining the current amplitude of each phase, a vector synthesis operation is required to calculate the comprehensive load current value. Because there is a 120-degree phase difference between the three phases of AC, the three amplitudes cannot be simply added together. The phase relationship needs to be considered for vector calculation. Specifically, the current amplitude of phase A is set as 0 degrees as the reference phase angle, phase B lags by 120 degrees, and phase C lags by 240 degrees. Then, the current amplitude of each phase is decomposed into real and imaginary parts. The real part of phase A is 50.6 multiplied by cos(0) equals 50.6, and the imaginary part is 50.6 multiplied by sin(0) equals 0. The real part of phase B is 49.8 multiplied by cos(-120 degrees) equals -24.9, and the imaginary part is 49.8 multiplied by sin(-120 degrees) equals -43. 1. The real part of phase C is 51.2 multiplied by cos(-240 degrees) equals -25.6, and the imaginary part is 51.2 multiplied by sin(-240 degrees) equals 44.3. Adding the real parts of the three phases together, we get 50.6 - 24.9 - 25.6 equals 0.1. Adding the imaginary parts together, we get 0 - 43.1 + 44.3 equals 1.2. Finally, we use the square root of the sum of the squares of the real and imaginary parts to calculate the comprehensive load current value, which is √(0.1² + 1.2²), approximately 1.2 amperes. This value reflects the degree of imbalance of the three-phase load. If the three phases are completely balanced, the comprehensive load current value will be close to 0. The larger the value, the more serious the imbalance. The comprehensive load current value is calculated in this way at every moment during the monitoring process.

[0114] After calculating a series of comprehensive load current values, they need to be time-series arranged according to the acquisition timestamps. This is because each data point in the three-phase current data carries a timestamp of its acquisition time. For example, the first data point's timestamp is 14:23:05.000, indicating it was acquired at 2:23:05 PM, and the second is 14:23:05.100, which is 100 milliseconds later. The system will arrange all the comprehensive load current values ​​in a column from morning to night according to their timestamps, resulting in the current time-series data points. However, in actual acquisition, communication delays or processor overload may cause data to be missed at certain times. Therefore, equal-interval verification is also necessary. The verification method is to check the timestamp difference between two adjacent data points. Normally, this should be 100 milliseconds. If a 200-millisecond time difference is found between two points, it indicates a missing point. The system will insert a compensation data point at this location, using the average of the two points. For example, if the current value at 14:23:08.300 is 1.5 amps, and the next data point is 14:23:08.500 with a current value of 1.8 amps, then the missing point is 14:23:08.400. The system will insert a compensation point timestamp of 14:23:08.400 with a current value of (1.5 + 1.8) / 2, which equals 1.65 amps. After filling in all missing points, the output is the standard timing current sequence, in which the data points are strictly arranged at 100-millisecond time intervals without any missing data.

[0115] After obtaining the standard time-series current sequence, adjacent data points need to be connected by curves to form a load characteristic curve. This connection operation uses either linear interpolation or cubic spline interpolation algorithms. Linear interpolation is simpler, connecting adjacent points with straight line segments, but the resulting curve may have sharp angles at transitions and lack smoothness. Cubic spline interpolation generates smoother curves. Specifically, a cubic polynomial function is used to fit between every two adjacent data points. The coefficients of the polynomial are solved using boundary and continuity conditions. If the standard time-series current sequence has 3000 data points, the system will generate 2999 cubic polynomial curves connecting these points. The two ends of each curve coincide with the data points, while the middle section transitions smoothly. The final generated load characteristic curve has the horizontal axis representing time from the moment the data acquisition is triggered to 5 minutes later, and the vertical axis representing the comprehensive load current value. The curve clearly shows the change of load current over time. For example, if a C-phase incoming terminal experiences an abnormal temperature due to the start-up of the load-side equipment, the load characteristic curve will show that the comprehensive load current value suddenly rises from 1.2 amps to 5.8 amps in the few seconds after the equipment starts up, and then gradually falls back to 1.5 amps. This curve is a concrete implementation of the technical element of "obtaining the load characteristic curve" in the upper-level solution. Through a series of processing steps, including effective value calculation, vector synthesis, timing arrangement, equal interval verification, and curve interpolation, the transformation from the original three-phase current data to the visualized characteristic curve is completed.

[0116] In a specific embodiment, fluctuation characteristics are extracted from the load characteristic curve to obtain load fluctuation parameters, including:

[0117] Extreme point detection is performed on the load characteristic curve to obtain the peak point and valley point of the curve, and the difference between the peak point and valley point of the curve is calculated to obtain the current fluctuation amplitude.

[0118] The slope of the current fluctuation amplitude is calculated based on the peak point and valley point of the curve to obtain the current change rate, and the current change rate is statistically calculated to obtain the mean rate and the peak rate.

[0119] The load fluctuation parameters are obtained by integrating the current fluctuation amplitude, the average rate, and the peak rate.

[0120] Specifically, when extracting fluctuation features from a load characteristic curve, the first step is to detect extreme points to identify peak and trough points on the curve. The detection method involves iterating through each data point on the curve and comparing its value with its two adjacent points. If the current value at a point is greater than both the previous and next points, it is considered a peak point; conversely, if it is less than both, it is considered a trough point. For example, the combined load current at 14:23:12.500 is 5.8 amps. At the previous time 14:23:12... The value at 0.400 is 5.3 amps, and the value at the next time 14:23:12.600 is 5.5 amps. Since 5.8 is greater than both 5.3 and 5.5, this point is marked as the peak point of the curve. Similarly, at 14:23:15.200, the current value drops to 1.1 amps, while the values ​​before and after are 1.4 amps and 1.3 amps respectively. This point is the valley point of the curve. The system records the coordinates of all detected peak and valley points. For example, 23 peak points and 21 valley points were found within a 5-minute monitoring window.

[0121] After detecting the extreme points, the current fluctuation amplitude needs to be calculated. The calculation method is to subtract the current value at the valley point from the current value at the peak point of the curve. However, there's a problem here: the number of peak points and valley points may be different, requiring pairing for calculation. The pairing rule is to find peak and valley points that are adjacent in time. For example, the peak point of 5.8 amps at 14:23:12.500 is immediately followed by the valley point of 1.1 amps at 14:23:15.200. Pairing these two points together, the fluctuation amplitude is 5.8 - 1.1 = 4.7 amps. Then, at 14:23:12.500... At 3:18.700, another peak value of 6.2 amps appears. Paired with the previous valley value of 1.1 amps, the fluctuation range is calculated to be 6.2 - 1.1 = 5.1 amps. After calculating the fluctuation range for all paired peak and valley values, a set of fluctuation range values ​​is obtained. The largest value in this set is used as the representative value of the current fluctuation range. If the calculated fluctuation ranges are 4.7, 5.1, 3.8, 4.2 amps, etc., then the maximum value of 5.1 amps is taken, which reflects the most severe fluctuation of the load current during the monitoring period.

[0122] Next, we need to calculate the rate of change of current. This rate is based on the slope between the peak and trough points. The slope is calculated by dividing the current difference by the time difference. Taking the peak of 5.8 amps at 14:23:12.500 and the trough of 1.1 amps at 14:23:15.200 as an example, the current difference is 5.8 - 1.1 = 4.7 amps, and the time difference is 15.2 seconds - 12.5 seconds = 2.7 seconds. Therefore, the slope of this decline is 4.7 divided by 2.7, approximately 1.74 amps per second. This value is the rate of change of current for that segment, indicating that the current decreases by an average of 1.74 amps per second. The same method is used to calculate the rate of increase from the trough to the next peak point, for example, from the trough of 1.1 amps at 14:23:15.200 to the trough of 14:23:15.200. At 23:18.700, the peak current is 6.2 amps. The current difference is 6.2 - 1.1 = 5.1 amps. The time difference is 18.7 - 15.2 = 3.5 seconds. The rise rate is calculated to be 5.1 divided by 3.5, which is approximately 1.46 amps per second. After calculating the rates between all adjacent peak and valley points, statistical calculations are performed. These calculations include finding the average rate and the peak rate. The average rate is calculated by adding up all rate values ​​and dividing by the total number of values. If the calculated rates are 1.74, 1.46, 2.15, 1.83, and 1.92 amps per second, the sum is 9.1, which is divided by 5 to get the average rate of 1.82 amps per second. The peak rate is the largest of these rates. Here, the peak rate is 2.15 amps per second, which represents the segment where the load current changes the fastest.

[0123] After obtaining the three parameters—current fluctuation amplitude, average rate, and peak rate—they need to be integrated to form the load fluctuation parameters. This integration operation involves organizing these three values ​​according to a defined data structure. The data structure can be a three-field structure: the first field stores the current fluctuation amplitude (5.1 amps), the second field stores the average rate (1.82 amps per second), and the third field stores the peak rate (2.15 amps per second). Alternatively, these three values ​​can be concatenated into a string separated by a delimiter, such as "5.1|1.82|2.15", or formed into an array [5.1, 1.82, ...]. [2.15] Regardless of the form, the final output is the load fluctuation parameter, which is a specific expansion of the technical element of "obtaining the load fluctuation parameter" in the upper-level scheme. Through a set of processes such as extreme value detection, difference calculation, slope calculation, statistical analysis and parameter integration, the fluctuation characteristics of the load characteristic curve are extracted with several key values. For example, if a certain B-phase outgoing terminal has an abnormal temperature due to frequent start-stop of the motor on the load side, the extracted load fluctuation parameters show that the current fluctuation amplitude reaches 5.1 Amperes, the average rate is 1.82 Amperes per second, and the peak rate is 2.15 Amperes per second. These data can help determine that the abnormal temperature is caused by load impact rather than a fault in the terminal itself.

[0124] In a specific embodiment, the operating status of the power metering box is evaluated based on the load fluctuation parameters to obtain the equipment health status, including:

[0125] The load fluctuation parameters are divided into thresholds to obtain fluctuation level values, and the fluctuation level values ​​are statistically analyzed over time to obtain the fluctuation duration.

[0126] The fluctuation level value is weighted based on the duration of the fluctuation to obtain a load state index, and the power metering box is described based on the load state index to obtain the equipment health status.

[0127] Specifically, after obtaining the load fluctuation parameters, the operating status of the electricity metering box needs to be assessed. The first step is to determine the fluctuation level by classifying thresholds. The system has several pre-set threshold standards. If the three conditions of current fluctuation amplitude less than 3 amps, average rate less than 1 amp per second, and peak rate not exceeding 1.5 amps per second are met simultaneously, it is classified as Level 1, indicating slight fluctuation. If the current fluctuation amplitude is between 3 and 5 amps, the average rate is between 1 and 2 amps per second, and the peak rate is between 1.5 and 2.5 amps per second, it is classified as Level 2, indicating moderate fluctuation. If the current fluctuation amplitude exceeds 5 amps, the average rate exceeds 2 amps per second, or the peak rate exceeds 2.5 amps per second, it is classified as Level 3, indicating severe fluctuation. For example, the load fluctuation parameters extracted earlier show a current fluctuation amplitude of 5.1 amps, an average rate of 1.82 amps per second, and a peak rate of 2.15 amps per second. Because the fluctuation amplitude of 5.1 amps exceeds the 5 amp limit, this set of parameters is classified as Level 3. This level value is recorded as the fluctuation level value.

[0128] After determining the fluctuation level, it's necessary to calculate how long this level lasted. The method involves returning to the load characteristic curve and identifying the time periods when the current fluctuation exceeded the threshold. For example, if the current jumps from 1.2 amps to 5.8 amps at 14:23:12.500, entering a level 3 fluctuation state, and continues until 14:23:45.300 when the current drops back to 2.8 amps, becoming a level 2 fluctuation, then the duration of the level 3 fluctuation is 45.3 seconds minus 12.5 seconds, equaling 32.8 seconds. The system divides the load characteristic curve into several segments, each corresponding to a fluctuation level. Then, all time periods of the same level are summed to obtain the total duration of that level. If a level 3 fluctuation occurred three times within a 5-minute monitoring window, with durations of 32.8 seconds, 18.5 seconds, and 25.2 seconds respectively, the total is 76.5 seconds. This value is the fluctuation duration, reflecting the proportion of time occupied by severe fluctuations during the entire monitoring period.

[0129] After obtaining the duration of the fluctuation, a weighted calculation is performed to determine the load state index. The principle of the weighted calculation is that the higher the fluctuation level and the longer the duration, the greater the impact on the equipment. Therefore, different weight coefficients are assigned to different levels. The weight coefficient for level 1 fluctuation is set to 1.0, level 2 to 1.5, and level 3 to 2.0. The calculation formula is to multiply the fluctuation level value by the corresponding weight coefficient, then multiply by the fluctuation duration, and then divide by the total monitoring time to obtain the contribution value for that level. Taking the previous example, the contribution value of level 3 fluctuation is 3 multiplied by 2.0 and then multiplied by... Dividing 76.5 seconds by 300 seconds equals 1.53. Using the same algorithm to handle level 2 fluctuations, if a level 2 fluctuation lasts for 135 seconds, the contribution value is 2 multiplied by 1.5 multiplied by 135 divided by 300, which equals 1.35. A level 1 fluctuation lasting for 88.5 seconds has a contribution value of 1 multiplied by 1.0 multiplied by 88.5 divided by 300, which equals 0.295. Adding the contribution values ​​of the three levels together, we get the load status index, which is approximately 3.18 (1.53 + 1.35 + 0.295). The higher this index value, the more severe the impact of load fluctuations on equipment operation.

[0130] After calculating the load state index, the status of the energy metering box needs to be described based on this index. There are three levels of status description: a load state index less than 2.0 is described as a healthy state, indicating that the equipment is operating normally; an index between 2.0 and 4.0 is described as a sub-healthy state, indicating that the equipment is operating at a certain risk; and an index exceeding 4.0 is described as a faulty state, indicating that the equipment is operating abnormally and requires immediate attention. The previously calculated load state index of 3.18 falls within the range of 2.0 to 4.0, so the health status of the energy metering box is judged as sub-healthy. This judgment result will be displayed on the monitoring interface and stored in the database. At the same time, the system will record the specific reasons for the sub-healthy state, such as a certain C-phase incoming line terminal. The problem stemmed from loose wiring bolts, which increased contact resistance and caused localized overheating. Infrared thermal imaging detected the abnormal temperature and triggered current acquisition. Analysis revealed that the load current fluctuated drastically during motor start-up and shutdown, with a fluctuation amplitude of 5.1 amperes lasting for 76.5 seconds. After comprehensive evaluation, the condition was determined to be sub-healthy. This assessment process is a refined implementation of the "obtaining equipment health status" technical element in the upper-level solution. Through a series of steps including threshold division, duration statistics, weighted calculation, and status description, the conversion from load fluctuation parameters to health status rating is completed. Compared with simply looking at temperature data, this assessment method combined with load analysis can more accurately determine whether the temperature abnormality is caused by equipment failure or normal temperature rise due to load characteristics.

[0131] Please see Figure 2 , Figure 2 This is a schematic diagram of the framework of an embodiment of an energy metering box with temperature monitoring function according to this application. Figure 2As shown, the scanning module 21 is used to perform infrared thermal imaging scanning on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box. The analysis module 22 is used to perform load current analysis on the power metering box based on the three-phase current data to obtain a load characteristic curve, and to extract the fluctuation characteristics of the load characteristic curve to obtain load fluctuation parameters. The evaluation module 23 is used to evaluate the operating status of the power metering box based on the load fluctuation parameters to obtain the equipment health status, and to execute the corresponding operation and maintenance strategy based on the equipment health status.

Claims

1. A monitoring method for an energy metering box with temperature monitoring function, characterized in that, include: Infrared thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box. Load current analysis is performed on the power metering box based on three-phase current data to obtain the load characteristic curve, and the fluctuation characteristics of the load characteristic curve are extracted to obtain the load fluctuation parameters. The operating status of the power metering box is evaluated based on the load fluctuation parameters to obtain the equipment health status, and the corresponding operation and maintenance strategy is executed based on the equipment health status. Infrared thermal imaging scans were performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map, including: Thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain terminal thermal imaging data; Based on the terminal thermal imaging data, the temperature values ​​of the incoming and outgoing terminals are calculated to obtain the terminal temperature matrix. Based on the terminal temperature matrix, the temperature gradient analysis of the terminal thermal imaging data is performed to obtain the terminal temperature distribution map. Based on the terminal temperature matrix, temperature gradient analysis is performed on the terminal thermal imaging data to obtain the terminal temperature distribution map, including: Based on the terminal temperature matrix, the temperature difference between adjacent points of the terminal thermal imaging data is calculated to obtain a temperature gradient vector map. The gradient direction of the temperature gradient vector map is statistically analyzed to obtain a temperature change trend line. The temperature gradient vector map is filtered by gradient amplitude threshold to obtain high gradient regions, and the center point of the region is located based on the high gradient regions to obtain the boundary of the heat accumulation area. Based on the boundary of the heat accumulation area, the terminal thermal imaging data is overlaid and labeled to obtain a labeled temperature map. Based on the temperature change trend line, the labeled temperature map is overlaid with the trend direction to obtain the terminal temperature distribution map. The terminal temperature distribution map is obtained by superimposing the labeled temperature map on the temperature change trend line, including: The endpoints of the temperature change trend line are extracted to obtain a trend endpoint set, and the direction angle is calculated based on the trend endpoint set to obtain the trend direction vector; The labeled temperature map is aligned with the coordinate system based on the trend direction vector to obtain an alignment reference point, and the trend direction vector is projected into space based on the alignment reference point to obtain a trend projection layer. The trend projection layer and the labeled temperature map are overlaid to obtain a fused temperature map, and the boundary contour of the fused temperature map is enhanced to obtain an enhanced boundary map. Based on the enhanced boundary map, directional arrows are marked on the trend projection layer to obtain a directional annotation map, and the directional annotation map is rendered with color levels to obtain the terminal temperature distribution map.

2. The monitoring method for an energy metering box with temperature monitoring function according to claim 1, characterized in that, Based on the alignment reference point, the trend direction vector is spatially projected to obtain a trend projection layer, including: Based on the alignment reference point, the image size of the labeled temperature map is read to obtain the image boundary range, and the scaling factor of the trend direction vector is calculated based on the image boundary range to obtain the projection scale factor. Based on the projection scale factor, the starting point coordinates in the trend direction vector are scaled and translated to obtain the transformation starting point; and based on the projection scale factor, the ending point coordinates in the trend direction vector are scaled and translated to obtain the transformation ending point. A connection path is generated based on the transformation start point and the transformation end point to obtain a vector line segment path, and a line width parameter is assigned to the vector line segment path to obtain a vector line segment diagram; The vector line segment graph is configured with a transparency channel to obtain a transparent vector graphic, and the transparent vector graphic is encapsulated in a layer format to obtain the trend projection layer.

3. The monitoring method for an energy metering box with temperature monitoring function according to claim 1, characterized in that, When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box, including: When the terminal temperature distribution map shows abnormal temperature spots, mark the coordinate position of the abnormal temperature spots in the terminal temperature distribution map; Based on the coordinate position, a temperature anomaly trigger signal is generated and sent to the real-time monitoring system. Upon receiving the temperature anomaly trigger signal, the real-time monitoring system immediately starts collecting three-phase current data from the power metering box. The data is collected according to a preset sampling frequency and sampling duration to obtain the three-phase current data.

4. The monitoring method for an energy metering box with temperature monitoring function according to claim 1, characterized in that, Load current analysis of the power metering box is performed based on three-phase current data to obtain load characteristic curves, including: Based on the current of each phase in the three-phase current data, the effective value of the power metering box is calculated to obtain the current amplitude of each phase, and the current amplitude of each phase is vector synthesized to obtain the comprehensive load current value. Based on the acquisition timestamps in the three-phase current data, the comprehensive load current value is time-series arranged to obtain current time-series data points, and the current time-series data points are checked at equal intervals to obtain a standard time-series current sequence. The load characteristic curve is obtained by interpolating adjacent standard time-series current sequences.

5. The monitoring method for an energy metering box with temperature monitoring function according to claim 1, characterized in that, The load characteristic curve is subjected to fluctuation feature extraction to obtain load fluctuation parameters, including: Extreme point detection is performed on the load characteristic curve to obtain the peak point and valley point of the curve, and the difference between the peak point and valley point of the curve is calculated to obtain the current fluctuation amplitude. The slope of the current fluctuation amplitude is calculated based on the peak point and valley point of the curve to obtain the current change rate, and the current change rate is statistically calculated to obtain the mean rate and the peak rate. The load fluctuation parameters are obtained by integrating the current fluctuation amplitude, the average rate, and the peak rate.

6. The monitoring method for an energy metering box with temperature monitoring function according to claim 1, characterized in that, The operating status of the power metering box is evaluated based on the load fluctuation parameters to obtain the equipment health status, including: The load fluctuation parameters are divided into thresholds to obtain fluctuation level values, and the fluctuation level values ​​are statistically analyzed over time to obtain the fluctuation duration. The fluctuation level value is weighted based on the duration of the fluctuation to obtain a load state index, and the power metering box is described based on the load state index to obtain the equipment health status.

7. An energy metering box with temperature monitoring function, characterized in that, The method for monitoring an energy metering box with temperature monitoring function according to any one of claims 1 to 6 includes: The scanning module is used to perform infrared thermal imaging scanning on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map. When the terminal temperature distribution map shows abnormal temperature spots, the real-time monitoring system is triggered to collect the three-phase current data of the power metering box. The analysis module is used to perform load current analysis on the power metering box based on three-phase current data, obtain the load characteristic curve, and extract the fluctuation characteristics of the load characteristic curve to obtain the load fluctuation parameters. The evaluation module is used to evaluate the operating status of the power metering box based on the load fluctuation parameters, obtain the equipment health status, and execute the corresponding operation and maintenance strategy based on the equipment health status. Infrared thermal imaging scans were performed on the incoming and outgoing terminals of the power metering box to obtain a terminal temperature distribution map, including: Thermal imaging scans are performed on the incoming and outgoing terminals of the power metering box to obtain terminal thermal imaging data; Based on the terminal thermal imaging data, the temperature values ​​of the incoming and outgoing terminals are calculated to obtain the terminal temperature matrix. Based on the terminal temperature matrix, the temperature gradient analysis of the terminal thermal imaging data is performed to obtain the terminal temperature distribution map. Based on the terminal temperature matrix, temperature gradient analysis is performed on the terminal thermal imaging data to obtain the terminal temperature distribution map, including: Based on the terminal temperature matrix, the temperature difference between adjacent points of the terminal thermal imaging data is calculated to obtain a temperature gradient vector map. The gradient direction of the temperature gradient vector map is statistically analyzed to obtain a temperature change trend line. The temperature gradient vector map is filtered by gradient amplitude threshold to obtain high gradient regions, and the center point of the region is located based on the high gradient regions to obtain the boundary of the heat accumulation area. Based on the boundary of the heat accumulation area, the terminal thermal imaging data is overlaid and labeled to obtain a labeled temperature map. Based on the temperature change trend line, the labeled temperature map is overlaid with the trend direction to obtain the terminal temperature distribution map. The terminal temperature distribution map is obtained by superimposing the labeled temperature map on the temperature change trend line, including: The endpoints of the temperature change trend line are extracted to obtain a trend endpoint set, and the direction angle is calculated based on the trend endpoint set to obtain the trend direction vector; The labeled temperature map is aligned with the coordinate system based on the trend direction vector to obtain an alignment reference point, and the trend direction vector is projected into space based on the alignment reference point to obtain a trend projection layer. The trend projection layer and the labeled temperature map are overlaid to obtain a fused temperature map, and the boundary contour of the fused temperature map is enhanced to obtain an enhanced boundary map. Based on the enhanced boundary map, directional arrows are marked on the trend projection layer to obtain a directional annotation map, and the directional annotation map is rendered with color levels to obtain the terminal temperature distribution map.

Images