Electric energy metering box safety system, method and device based on multi-source data

Abstract

The application provides a kind of based on multi-source data's electric energy metering box safety system, method and device, system contains current feature acquisition subsystem, lightweight coupling subsystem and dispatch interaction subsystem.Method contains time-frequency spectrum feature automatic triggering and dynamic demarcation infrared thermal imaging scanning area, only to the physical node of current abnormal loop association carries out millisecond level high-resolution thermal field capture;Three-dimensional thermal topology vector and homologous current ripple feature carry out space-time alignment, generate thermal-electricity combined abnormal coefficient through lightweight coupling algorithm, comprehensively represent the correlation intensity of heating position, temperature rise rate and current disturbance, automatically remove irrelevant interference such as environmental temperature fluctuation;Based on thermal-electricity combined abnormal coefficient, independently generate the risk penetration index of grading, real-time drive and power grid dispatching system two-way communication;And accordingly reverse optimization current ripple trigger threshold.The application realizes the security protection upgrade from passive alarm to active prediction.

Description

Technical Field

[0001] This invention relates to the field of electrical equipment safety technology, and in particular to a safety system, method and device for an electrical metering box based on multi-source data. Background Technology

[0002] With the accelerated construction of smart grids, electricity metering boxes, as key nodes connecting the power supply network and users, directly impact grid stability and the quality of electricity supply for users. Currently, metering boxes face three major challenges: environmental complexity, high incidence of human-caused damage, and data silos. The industry urgently needs to build an integrated protection system that enables real-time sensing, intelligent analysis, and proactive defense.

[0003] Prior art 1, application number: CN202510285290.7, discloses a safety energy metering box for preventing electric shock, including a box body and an insulated mounting frame. The mounting frame includes a heat dissipation rack, which includes at least two rows of mounting rods. Spherical structures are connected in series on the mounting rods. Adjacent spherical structures on the same mounting rod are tangent, and spherical structures between adjacent mounting rods in the same row are tangent. The center heights of adjacent spherical structures in adjacent rows of mounting rods are different. The box body has heat dissipation grooves. The spheres are slidably mounted on the mounting rods, and a pressure block is helically mounted at one end of each mounting rod. A mounting seat is mounted on the heat dissipation rack. The mounting seat includes a mounting ring and a mounting plate. The mounting ring has through holes and clearance grooves. The mounting seat is mounted on the mounting rods in the same row, and a linkage rod is connected to the mounting rods in the same row. The linkage rod and the base plate are in sliding contact. Although this allows for effective heat dissipation and flexible position adjustment of the electronic components inside the energy metering box, it only optimizes the physical heat dissipation structure and does not solve the problem of active monitoring of electrical anomalies such as current disturbances and localized overheating.

[0004] Prior art 2, application number: CN202411451683.2, discloses an intelligent safety protection electricity metering box, including a metering box shell, a protective shell, a bottom shell, a top shell, and a tilting safety protection mechanism; wherein the tilting safety protection mechanism includes a sliding frame, a concave block, a support shaft, a protective sleeve, a tilting groove strip, a screw, a hinged sleeve block, and a connecting shaft; a tilting groove strip is installed on one side of the inner wall of the bottom shell, and a linkage safety protection mechanism is provided on one side of the tilting groove strip; wherein a distributed detection component is provided on the other side of the tilting groove strip. Although the tilting safety protection mechanism can promptly reinforce the internal tilting of the protective shell, thus improving the intelligent safety protection of the electricity metering box and solving the problem of difficulty in timely detection of electricity theft, which makes it difficult to promptly reinforce the outer shell of each electricity metering box, resulting in poor intelligent safety protection; however, it only addresses physical anti-theft and lacks intelligent diagnostic capabilities for internal electrical faults such as overload and poor contact.

[0005] Existing technology three, application number: CN202310006189.4, discloses a safe energy metering box. It uses two parallel buffer plates arranged around the outer perimeter of the inner box. When a buffer plate is impacted, the impacted plate depresses against its contacting return spring and displaces along a guide rail. Because the buffer plate on the other side remains positioned under the limiting action of its corresponding limiting part, the guide rod can effectively pass through the impacted buffer plate, ensuring the buffer plate's effectiveness in buffering external impact forces. After the impact disappears, the buffer plate gradually returns to its original position under the restoring force of the return spring, re-establishing a sealing effect on the inner box. Therefore, it not only forms a protective barrier on the outer perimeter of each meter through the buffer plates, but also, with the buffer plates and guide rods spaced apart from the inner box, prevents the metering boxes from being easily damaged by the inner box's swaying. Although the buffer plates reduce damage to themselves and extend their service life through buffer displacement, they only provide mechanical protection and cannot identify early signs of electrical hazards such as high-frequency oscillations and abnormal temperature rises.

[0006] Currently, existing technologies 1, 2, and 3 suffer from difficulties in early identification of electrical anomalies, low fault location accuracy, and rigid operation and maintenance response strategies. Therefore, this invention provides a safety system, method, and device for power metering boxes based on multi-source data. Summary of the Invention

[0007] To address the aforementioned technical problems, this invention provides a safety system for an energy metering box based on multi-source data, comprising: The lightweight coupling subsystem is used to deconstruct the data of the scanned area into a three-dimensional thermal topology vector in real time within the edge node; the three-dimensional thermal topology vector is spatiotemporally aligned with the ripple characteristics of the same source current, and a thermal-electric joint anomaly coefficient is generated through the lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance. The scheduling interaction subsystem is used to autonomously generate a graded risk penetration index based on the thermal-electric joint anomaly coefficient, and drive bidirectional communication with the power grid scheduling system in real time: when the penetration is high, it actively pushes fault location maps; when the penetration is low, it only receives the operation strategy issued by the scheduling side, and optimizes the trigger threshold of current ripple accordingly.

[0008] Optional, a scheduling interaction subsystem includes: The anomaly coefficient processing component is used to process the combined thermal-electric anomaly coefficient through the heat capacity mapping of the conduction path, and to call the preset path heat capacity parameters in the three-dimensional vector field; the path heat capacity parameters are processed by the anomaly coefficient weighting, and the original anomaly coefficient is multiplied by the reciprocal of the path heat capacity to generate a heat capacity weighted anomaly coefficient that suppresses the interference of high heat capacity paths. The pattern matching processing component is used to process the heat capacity weighted anomaly coefficient through historical failure mode matching. The matching results are then processed by permeability intensity calculation. Based on the standard deviation multiple of the current anomaly coefficient from the historical benchmark value, a single-path permeability intensity value is generated. The permeability intensity is then output based on the path length in the three-dimensional vector field, which decreases exponentially. The indicator output component is used to process the single-path permeability through the topological expansion of the transmission path, mapping the linear path permeability to three-dimensional spatial coordinates; extracting the gradient extreme points of permeability spatial variation; aggregating the weighted sum of permeability of all paths within the radius centered on the gradient extreme points; and outputting high, medium, and low risk permeability indicators according to the preset physical range of the superimposed value.

[0009] Optional, indicator output components, including: The integral processing sub-component is used to process the spatial permeability distribution data through heat flow direction constraints, and to limit the principal axis of gradient calculation using the heat flow direction data of the three-dimensional thermal topology vector; to generate a heat conduction gradient field; the heat conduction gradient field is processed through potential field integral transformation, and the gradient field is inversely integrated along the heat flow path to generate a risk potential field characterizing the intensity of risk accumulation; The verification processing sub-component is used to process the risk potential field data through potential energy surface modeling to construct a three-dimensional continuous potential energy distribution surface; the potential energy distribution surface is processed by saddle point detection to identify critical points on the surface; the critical points are processed by heat flow path verification to filter points located on the preset conduction path of the three-dimensional vector field and output the coordinates of the gradient extreme points. The aggregation processing sub-component maps the gradient extreme point coordinates to the corresponding physical conduction paths in the 3D vector field after the conduction path association processing. The associated physical conduction paths are processed by the neighborhood thermal resistance network construction processing, connecting all physical conduction paths with heat flow within the radius centered on the polar coordinate point. The neighborhood thermal resistance network is processed by path thermal resistance weighting processing, calculating the weighting coefficient of each path according to the thermal resistance parameters of the material library. The weighting coefficients are processed by permeability aggregation processing, which performs thermal resistance weighted summation on the single-path permeability of all paths in the neighborhood to generate a spatial aggregated permeability value.

[0010] Optional, the aggregation processing sub-component includes: The dimensional integration module is used to process the path thermal resistance data in the neighborhood thermal resistance network through material conductivity mapping and call the thermal conductivity parameters of the corresponding conductors in the material library; the thermal conductivity parameters are processed through path conduction efficiency, and the thermal conductivity value is integrated with the path cross-sectional area and length to generate the path conduction efficiency coefficient. The proportion conversion module is used to process the path conduction efficiency coefficient after neighborhood heat flow balance and calculate the node heat flow conservation equation based on the connection topology of each path in the thermal resistance network; the node heat flow balance solution is processed by efficiency weight normalization and the conduction efficiency coefficient is converted into weight coefficient according to the proportion of the path in the neighborhood. The inertia compensation module is used to process the heat flow weight coefficient through permeability-heat flow coupling, multiplying the single-path permeability with the weight coefficient to characterize the contribution of that path to the risk propagation in the neighborhood; the coupled contribution is aggregated through heat flow direction processing, summing the contributions of all paths in the neighborhood along the main heat flow direction defined by the three-dimensional thermal topology vector; the summation result is processed through thermal inertia compensation, and the time accumulation effect of the aggregated value is corrected according to the specific heat capacity parameter of the path material to generate a spatial aggregated permeability value.

[0011] Optional, percentage conversion module, including: The proportional allocation submodule is used to establish a physical balance at the intersection of thermal resistance network paths by processing the path conduction efficiency coefficient through heat flow conservation in the junction area, ensuring that the total heat flow input equals the total heat flow output. After the heat flow conservation balance is processed by conduction capacity allocation, the heat flow in the junction area is allocated according to the relative proportion of the conduction efficiency coefficient of each path, generating the heat flow ratio of each path. The convergence processing submodule is used to process the path heat flow ratio value through multi-level convergence coupling, and to transfer the heat flow distribution results of adjacent convergence zones according to the thermal resistance network topology hierarchy. The hierarchical data transfer is processed by heat flow direction consistency verification to ensure that the heat flow direction of each path is consistent with the main direction defined by the three-dimensional thermal topology vector. The verified heat flow ratio is processed by network balance convergence, and iteratively adjusted until the heat flow distribution of the entire network satisfies the conservation conditions of all convergence zones, generating the node heat flow balance solution. The normalization submodule is used to normalize the proportion of path heat flow in the node heat flow balance solution, converting the heat flow proportion of each path in the neighborhood into a value in the range of [0,1]. The normalized proportion is then processed by material thermal hysteresis compensation, and the influence of heat flow transfer delay is corrected according to the thermal diffusion rate parameter of the material library to generate heat flow weight coefficient.

[0012] Optional, proportional allocation submodule, includes: The ratio processing unit is used to process the path conduction efficiency ratio in the heat flow conservation equilibrium to obtain the ratio of the conduction efficiency coefficient of each path in the intersection area to the maximum efficiency coefficient; the path conduction efficiency ratio is processed by the heat flow allocation weight generation process, and the square root of the ratio is mapped to the initial value of the allocation weight to generate the normalized allocation weight. The condition calibration unit is used to normalize the allocation weights. After the conservation condition calibration process, the weight values ​​are adjusted according to the total input heat flux in the junction area to ensure that the sum of the allocation weights of each path is equal to 1. The calibration weights are then processed by material interface effect correction, which calls the thermal reflection coefficients of different conductor connection interfaces in the material library to correct the heat flux loss at the interface. The corrected weights are then processed by heat flux ratio calculation, which multiplies the total input heat flux by the weight values ​​of each path to generate the physical allocation ratio. The density conversion unit is used to distribute the input heat flow to each output path according to the output path heat flow synthesis process. The distributed heat flow value is processed by path heat flow density conversion and divided by the cross-sectional area of ​​the path to obtain the heat flow per unit area. The heat flow per unit area is normalized by the proportion value to calculate the percentage of heat flow of each path to the total output heat flow of the junction area, and generate the path heat flow proportion value.

[0013] Optional, condition calibration unit, comprising: The path allocation subunit is used to decompose the normalized weights into independent allocation factors corresponding to specific output paths after the corrected weight data has undergone path independence calculation. The independent allocation factors are then bound to the total heat flux to match the physical dimensions of each factor with the total input heat flux, generating dimensional path allocation factors. The conservation verification subunit is used to process the dimensional path allocation factor by multiplying it by the total heat flux input in the junction area to generate the original heat flux allocation value. The original heat flux allocation value is then processed by interface reflection loss compensation, which deducts the interface energy loss based on the interface thermal reflection coefficient of the material library. The compensated heat flux value is then processed by path heat flux conservation verification to generate the path heat flux value. The constraint processing subunit is used to process the path heat flux value through the path heat flux density, divide it by the maximum allowable flow rate of the path to convert it into a dimensionless proportional value; the dimensionless proportional value is then processed by the proportional range constraint to linearly map the proportional value to the [0,1] physical interval to generate the physical allocation ratio.

[0014] Optionally, it also includes a current feature acquisition subsystem for continuously analyzing the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics automatically trigger and dynamically delineate the scanning area of ​​infrared thermal imaging, and only perform millisecond-level high-resolution thermal field capture on the physical nodes associated with the current abnormal circuit.

[0015] This invention provides a safety method for electricity metering boxes based on multi-source data, comprising the following steps: The time-spectrum characteristics of the current ripple inside the power metering box are continuously analyzed. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically defined. Only the physical nodes associated with the current abnormal circuit are captured with millisecond-level high-resolution thermal field. Within the edge nodes, the data of the scanned area is deconstructed into a three-dimensional thermal topology vector in real time; the three-dimensional thermal topology vector is spatiotemporally aligned with the ripple characteristics of the same source current, and a thermal-electric joint anomaly coefficient is generated through a lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance. Based on the combined thermal and electrical anomaly coefficient, a graded risk penetration index is autonomously generated to drive real-time bidirectional communication with the power grid dispatch system: when the penetration is high, fault location maps are actively pushed; when the penetration is low, only the operation strategy issued by the dispatch side is received, and the trigger threshold of current ripple is optimized accordingly.

[0016] The present invention provides a safety device for an energy metering box based on multi-source data, comprising: an energy metering box door, a display screen, a control box, and a box body; The front left side of the enclosure has an electricity metering box door, the right side of the electricity metering box door has multiple displays, and a control box is installed on one side of the electricity metering box door. The control box contains an integrated circuit board, which stores programs and instructions related to the current characteristic acquisition subsystem, the lightweight coupling subsystem, and the scheduling interaction subsystem.

[0017] This invention achieves accurate monitoring and proactive defense of the safety status of power metering boxes through multi-source data fusion and intelligent analysis. The current characteristic acquisition subsystem uses time-spectrum analysis to lock onto the high-frequency oscillation characteristics of abnormal currents, dynamically guiding the infrared thermal imaging system to perform targeted scanning. This transforms traditional full-domain temperature measurement into millisecond-level thermal field capture of abnormal loop-related nodes, improving detection efficiency while avoiding invalid data collection. The lightweight coupling subsystem constructs a joint analysis model of thermal and electrical dual-physics fields. By aligning the three-dimensional thermal topology vector with the spatiotemporal data of current ripple, it establishes a quantitative correlation between temperature rise characteristics and current disturbances (thermal-electric joint anomaly coefficient), effectively eliminating environmental interference and reducing the false alarm rate compared to single-parameter monitoring. The scheduling interaction subsystem achieves autonomous quantification of risk levels, forming a graded response mechanism: proactively pushing complete diagnostic data including fault location maps during high penetration; implementing preventative maintenance through dynamic threshold adjustment during low penetration; enabling the power grid dispatching system to adapt different handling strategies according to risk levels, optimizing the efficiency of operation and maintenance resource allocation.

[0018] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0019] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0020] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings: Figure 1This is a block diagram of the power metering box safety system based on multi-source data in Embodiment 1 of the present invention; Figure 2 This is a schematic diagram of the safety system for an energy metering box based on multi-source data in Embodiment 1 of the present invention; Figure 3 This is a block diagram of the current characteristic acquisition subsystem in Embodiment 2 of the present invention; Figure 4 This is a block diagram of the lightweight coupling subsystem in Embodiment 5 of the present invention; Figure 5 This is a block diagram of the scheduling interaction subsystem in Embodiment 8 of the present invention; Figure 6 This is a flowchart of the power metering box security method based on multi-source data in Embodiment 14 of the present invention; Figure 7 This is a structural diagram of the safety device for the power metering box based on multi-source data in Embodiment 15 of the present invention. Detailed Implementation

[0021] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0022] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to limit the embodiments of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more associated listed items.

[0023] In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application. In the description of this application, it should be understood that the terms "first," "second," "third," etc., are used only to distinguish similar objects and are not necessarily used to describe a specific order or sequence, nor should they be construed as indicating or implying relative importance. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0024] Example 1: As Figure 1 As shown, this embodiment of the invention provides a safety system for an energy metering box based on multi-source data, comprising: The current characteristic acquisition subsystem is used to continuously analyze the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically defined. Only the physical nodes associated with the current abnormal circuit are captured with millisecond-level high-resolution thermal field. The lightweight coupling subsystem is used to deconstruct the data of the scanned area into a three-dimensional thermal topology vector in real time within the edge node; the three-dimensional thermal topology vector is spatiotemporally aligned with the ripple characteristics of the same source current, and a thermal-electric joint anomaly coefficient is generated through the lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance, and automatically eliminate irrelevant interference such as ambient temperature fluctuations. The scheduling interaction subsystem is used to autonomously generate a graded risk penetration index based on the thermal-electric joint anomaly coefficient, and drive bidirectional communication with the power grid scheduling system in real time: when the penetration is high, it actively pushes fault location maps; when the penetration is low, it only receives the operation strategy issued by the scheduling side, and optimizes the trigger threshold of current ripple accordingly.

[0025] The working principle and beneficial effects of the above technical solution are as follows: The current feature acquisition subsystem of this embodiment is used to continuously analyze the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically delineated. Only the physical nodes associated with the current abnormal loop are captured at a high-resolution thermal field level in milliseconds. The lightweight coupling subsystem is used to deconstruct the data of the scanning area into a three-dimensional thermal topology vector in real time within the edge node. The three-dimensional thermal topology vector is spatiotemporally aligned with the homogeneous current ripple features. The lightweight coupling algorithm generates a thermal-electric joint anomaly coefficient, which comprehensively characterizes the correlation strength between the heating location, the temperature rise rate and the current disturbance, and automatically eliminates irrelevant interference such as ambient temperature fluctuations. The scheduling interaction subsystem is used to autonomously generate a graded risk penetration index based on the thermal-electric joint anomaly coefficient, and drive bidirectional communication with the power grid dispatching system in real time: when the penetration is high, the fault location map is actively pushed; when the penetration is low, only the operation strategy issued by the dispatching side is received, and the trigger threshold of the current ripple is optimized accordingly (the specific principle is as follows). Figure 2(As shown). The above scheme achieves accurate monitoring and proactive defense of the safety status of the power metering box through multi-source data fusion and intelligent analysis. The current characteristic acquisition subsystem locks the high-frequency oscillation characteristics of abnormal current through time-spectrum analysis, dynamically guiding the infrared thermal imaging system to perform targeted scanning, transforming traditional full-domain temperature measurement into millisecond-level thermal field capture of abnormal loop associated nodes, improving detection efficiency while avoiding invalid data collection. The lightweight coupling subsystem constructs a joint analysis model of thermal-electric dual physical fields, establishing a quantitative correlation between temperature rise characteristics and current disturbances (thermal-electric joint anomaly coefficient) through the spatiotemporal alignment of three-dimensional thermal topology vectors and current ripple, effectively eliminating environmental interference and reducing the false alarm rate compared to single-parameter monitoring. The dispatch interaction subsystem realizes autonomous quantification of risk levels, forming a hierarchical response mechanism: proactively pushing complete diagnostic data including fault location maps when penetration is high; implementing preventive maintenance through dynamic threshold adjustment when penetration is low; enabling the power grid dispatch system to adapt different handling strategies according to risk levels, optimizing the efficiency of operation and maintenance resource allocation.

[0026] In summary, this embodiment constructs a technical chain of feature capture, coupled analysis, and policy closed loop, achieving a security protection upgrade from passive alarm to proactive prediction without excessively increasing the data transmission burden (relying on edge computing).

[0027] Example 2: Figure 3 As shown, based on Embodiment 1, the current characteristic acquisition subsystem provided in this embodiment of the invention includes: The focused feature extraction component is used to adaptively decompose the raw current signal flow in the power metering box through a multi-scale window to obtain the time-frequency energy distribution matrix; the time-frequency energy distribution matrix is ​​then subjected to transient oscillation feature extraction to extract the high-frequency band energy accumulation feature; The dynamic delineation component is used to identify specific high-frequency oscillation modes by analyzing the abnormal pattern matching of high-frequency band energy accumulation characteristics; the specific high-frequency oscillation modes are then mapped by physical node association, and based on preset loop topology data, the coordinate set of physical nodes that are only associated with abnormal loops is dynamically delineated. The scan range generation component is used to optimize the physical node coordinate set through the thermal imaging area to generate a minimized scan range; the minimized scan range is driven by a high-speed thermal sensor to perform millisecond-level high-resolution thermal field data capture.

[0028] The working principle and beneficial effects of the above technical solution are as follows: The focusing feature extraction component in this embodiment is used to adaptively decompose the original current signal flow in the power metering box through a multi-scale window to obtain a time-frequency energy distribution matrix; the time-frequency energy distribution matrix is ​​then processed by transient oscillation feature extraction to extract high-frequency band energy accumulation features; the dynamic delineation component is used to identify specific high-frequency oscillation modes by performing abnormal pattern matching analysis on the high-frequency band energy accumulation features; the specific high-frequency oscillation modes are then mapped by physical node association, and based on preset loop topology data, the coordinate set of physical nodes that are only associated with abnormal loops is dynamically delineated; the scanning range generation component is used to optimize the physical node coordinate set through thermal imaging area to generate a minimized scanning range; the minimized scanning range is driven by a high-speed thermal sensor to perform millisecond-level high-resolution thermal field data capture. The current feature acquisition subsystem of the above scheme achieves a complete process from raw current signal to precise location of abnormal loop through multi-module collaborative operation. First, the current signal is decomposed at multiple scales to extract high-frequency energy accumulation patterns that reflect abnormal characteristics. Then, through abnormal pattern matching analysis, fault signals with specific oscillation characteristics are identified and accurately correlated to specific physical nodes using loop topology data. Finally, thermal imaging area optimization technology compresses the detection range to the minimum necessary area, guiding thermal sensing equipment to efficiently complete high-resolution thermal field data acquisition. The entire process achieves closed-loop processing from electrical signal feature extraction to precise physical spatial location, providing a high-efficiency and high-accuracy technical solution for abnormal loop detection. The decomposition results of the raw signal stream are directly used to construct a time-frequency energy distribution matrix. The transient characteristics of this matrix are directly input into abnormal pattern matching analysis. The matching result, i.e., the specific high-frequency oscillation pattern, directly drives the physical node association mapping. The mapping output directly defines the input for thermal imaging area optimization, and the optimized output directly controls the high-speed thermal sensor for capture. All steps are based on the inherent correlation of current abnormal characteristics, avoiding full-area scanning and environmental interference.

[0029] Example 3: Based on Example 2, the scanning range generation component provided in this embodiment of the invention includes: The matrix construction sub-component is used to model the spatial adjacency relationship between the physical node coordinate set and the pre-set loop topology data, and construct a spatial constraint relationship matrix containing the conduction paths between abnormal nodes; the spatial constraint relationship matrix is ​​compensated for by the thermal resistance coefficient of the material, and combined with the conductor material library data to predict the maximum thermal diffusion boundary; The vector field synthesis sub-component is used to dynamically focus the predicted thermal diffusion boundary results and synthesize a three-dimensional vector field covering the potential heat conduction path. The region elimination sub-component is used to compress and reorganize the 3D vector field through the scan path, eliminate spatially overlapping regions, and generate a minimized scan range that only contains the key monitoring areas.

[0030] The working principle and beneficial effects of the above technical solution are as follows: The matrix construction sub-component in this embodiment is used to model the spatial adjacency relationship between the physical node coordinate set and the preset loop topology data, constructing a spatial constraint relationship matrix containing the conduction paths between abnormal nodes; the spatial constraint relationship matrix is ​​compensated for by the material thermal resistivity coefficient, and combined with conductor material library data to predict the maximum thermal diffusion boundary; the vector field synthesis sub-component is used to synthesize a three-dimensional vector field covering potential heat conduction paths by dynamically focusing the predicted thermal diffusion boundary results; the region elimination sub-component is used to compress and reorganize the three-dimensional vector field through scanning path compression, eliminating spatially overlapping areas and generating a minimized scanning range containing only key monitoring areas. The above solution achieves precise positioning and efficiency optimization for thermal conduction anomaly monitoring through modular collaboration: the matrix construction sub-component establishes a mathematical model of the heat conduction path by fusing physical topology and material property data, quantitatively reflecting the spatial constraint relationship and conduction limit during the thermal diffusion process. Based on the aforementioned boundary prediction, the vector field synthesis sub-component uses dynamic focusing technology to construct a three-dimensional conduction probability distribution, transforming discrete node data into potential heat flow trajectories in continuous space. The region elimination sub-component reduces the scanning range by spatial redundancy removal while maintaining monitoring integrity. The final output of the minimized critical region significantly reduces the computational load and time cost of the detection system.

[0031] In summary, this embodiment achieves end-to-end transformation from raw topology data to optimized monitoring areas. Its core value lies in extracting high-value monitoring targets from complex heat conduction systems through multi-stage data dimensionality reduction and spatial mapping, providing a decision-making basis for the efficient deployment of subsequent detection equipment. Each module forms a progressive processing chain, with each preceding output serving as the input for subsequent operations, ultimately achieving a balance between accuracy and efficiency. The spatial constraint matrix is ​​directly constructed from the physical node coordinate set and pre-set loop topology data; heat diffusion boundary prediction couples the spatial constraint matrix with the conductor material library; the three-dimensional scanning vector field is directly converted from the heat diffusion boundary prediction results using a dynamic focusing algorithm; and minimizing the scanning range is directly generated through path compression of the three-dimensional scanning vector field. The entire process, through dual constraints of conduction path modeling and heat diffusion prediction, transforms the spatial correlation of abnormal nodes into a precise scanning area, avoiding redundant operations based on fixed partitions or full-range scanning in existing technologies.

[0032] Example 4: Based on Example 3, the vector field synthesis sub-component provided in this embodiment of the invention includes: The heat flow principal direction field generation module is used to process the spatial temperature extreme value distribution data in the heat diffusion boundary prediction results through directional gradient decomposition to separate the temperature change rate components of each axis in three-dimensional space. The temperature change rate components of each axis in three-dimensional space are then processed by the principal conduction direction filtering process to extract the spatial vector with the largest temperature gradient change as the dominant heat flow direction. The dominant heat flow direction is then processed by neighborhood coherence verification to connect the direction vectors of adjacent spatial points to form a continuous conduction path, thus generating the heat flow principal direction field. The focused intensity distribution generation module uses the thermal resistance coupling processing of the conduction path direction data in the main heat flow direction field, combined with the thermal conductivity parameters of the conductor material library, to calculate the heat flow resistance coefficient at different spatial locations. The heat flow resistance coefficient is then processed by directional weight transformation, mapping the resistance coefficient inversely to the scanning focused intensity weight. The scanning focused intensity weight is then processed by spatial interpolation enhancement, generating a continuous weight distribution at non-measurement point locations based on the intensity gradient of adjacent conduction paths, thus forming an anisotropic focused intensity distribution. The scanning vector field generation module is used to process the spatial weight data of the anisotropic focusing intensity distribution through vector basis construction, and to establish a local coordinate system based on the main direction of heat flow at the coordinates of the abnormal nodes. The local coordinate system is then subjected to intensity-direction binding processing, and the focusing intensity weight is projected onto the three orthogonal axial components of the coordinate system. The three orthogonal axial components are then subjected to spatial vector superposition processing, and the conduction path direction and focusing intensity are fused to generate a three-dimensional scanning vector with magnitude and direction, forming the initial three-dimensional scanning vector field. The three-dimensional vector field generation module is used to process discrete vector data in the initial three-dimensional scan vector field through path breakpoint detection to identify spatial locations where vector direction changes abruptly or magnitude drops sharply in the conduction path. The spatial breakpoint locations are then processed by heat flow continuity reconstruction, and transition compensation vectors are generated based on the directional trends and intensity attenuation rates of adjacent vectors. The transition compensation vectors are then processed by vector field smoothing convergence to eliminate local oscillations and ensure continuous change in vector direction in the conduction path, ultimately generating a three-dimensional vector field covering the potential heat conduction path.

[0033] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the heat flow main direction field generation module uses directional gradient decomposition to separate the temperature change rate components along each axis of the three-dimensional space from the spatial temperature extreme value distribution data in the heat diffusion boundary prediction results; the temperature change rate components along each axis of the three-dimensional space are then filtered by the main conduction direction to extract the spatial vector with the largest temperature gradient change as the dominant heat flow direction; the dominant heat flow direction undergoes neighborhood coherence verification to connect the direction vectors of adjacent spatial points to form a continuous conduction path, generating the heat flow main direction field; the focusing intensity distribution generation module uses material thermal resistance coupling processing for the conduction path direction data in the heat flow main direction field, combined with the thermal conductivity parameters of the conductor material library, to calculate the heat flow resistance coefficient at different spatial locations; the heat flow resistance coefficient undergoes directional weight conversion processing to inversely map the resistance coefficient to the scanning focusing intensity weight; the scanning focusing intensity weight undergoes spatial interpolation enhancement processing to generate a continuous weight distribution at non-measurement point locations based on the intensity gradient of adjacent conduction paths, forming an anisotropic focusing intensity distribution; the scanning vector field generation module uses vector basis processing for the spatial weight data of the anisotropic focusing intensity distribution... The process involves establishing a local coordinate system at the coordinates of abnormal nodes, with the main direction of heat flow as the reference. This local coordinate system undergoes intensity-direction binding, projecting the focusing intensity weight onto the three orthogonal axial components of the coordinate system. These three orthogonal axial components are then subjected to spatial vector superposition, fusing the conduction path direction and focusing intensity to generate a three-dimensional scanning vector with magnitude and direction, forming an initial three-dimensional scanning vector field. A three-dimensional vector field generation module is used to detect path breakpoints in the discrete vector data within the initial three-dimensional scanning vector field, identifying spatial locations where vector direction changes abruptly or magnitude drops sharply in the conduction path. The spatial breakpoint locations are then processed by heat flow continuity reconstruction, generating transition compensation vectors based on the directional trends and intensity attenuation rates of adjacent vectors. These transition compensation vectors undergo vector field smoothing convergence processing to eliminate local oscillations and ensure continuous change in the vector direction of the conduction path, ultimately generating a three-dimensional vector field covering potential heat conduction paths. The above scheme achieves three-dimensional dynamic modeling and scanning optimization of the heat conduction path through multi-level processing: the heat flow main direction field generation module transforms temperature gradient data into a continuous heat flow dominant direction field, clarifying the main conduction trend of heat diffusion; the focused intensity distribution generation module combines material properties to quantify heat flow resistance and convert it into scanning weights, forming an anisotropic detection intensity distribution; the scanning vector field generation module fuses direction and intensity to construct an initial three-dimensional vector field, ensuring that the heat flow characteristics of key nodes are accurately characterized; the three-dimensional vector field generation module optimizes path continuity, eliminates abrupt changes and breakpoints, and forms a smooth and physically reasonable complete vector field. This embodiment transforms the heat diffusion prediction results into an executable three-dimensional vector field through gradient analysis, material coupling, vector construction, and path optimization, preserving the core conduction characteristics of heat flow while providing a high-precision spatial data foundation for subsequent region removal.Each module undergoes progressively refined processing to ensure that the final vector field possesses accuracy in direction, reasonable intensity, and path continuity. Through a five-stage progressive process—gradient decomposition, direction filtering, thermal resistance coupling, vector construction, and discontinuity compensation—the static boundary of the thermal diffusion prediction is transformed into a dynamic scanning vector field. The principal direction of heat flow determines the vector direction, the material's thermal resistance characteristics determine the vector magnitude, and path discontinuity compensation ensures the continuity of physical conduction, achieving a closed-loop generation of the entire chain: prediction boundary, conduction direction scanning intensity, and path coherence.

[0034] Example 5: Figure 4 As shown, based on Embodiment 1, the lightweight coupling subsystem provided in this embodiment of the invention includes: The 3D thermal topology vector generation component is used to process the 3D vector field data obtained by infrared thermal imaging scanning through vector-temperature field conversion, mapping the directional attributes and modulus intensity of the scanning vector to a spatial temperature gradient distribution. The spatial temperature gradient distribution is then verified through heat flow path verification, using the pre-defined heat conduction direction of the 3D vector field to constrain the actual temperature gradient direction, generating a purified temperature field free of non-correlated thermal noise. The purified temperature field is then processed through dynamic thermal topology modeling, extracting the temperature rise rate and heat flow direction of key nodes along the predefined conduction path of the 3D vector field to generate a 3D thermal topology vector. The spatiotemporal coupling feature set generation component is used to process the heat flow direction data of the three-dimensional thermal topology vector through current phase coupling, aligning the current ripple oscillation period with the heat flow direction change period; the aligned current features are then processed by ripple-heat flow correlation extraction, binding the high-frequency amplitude and spectral entropy of the current at the corresponding time to the spatial coordinate position of the three-dimensional thermal topology vector; the bound data are then processed by spatiotemporal feature matrix reconstruction, arranging the current-thermal field parameters according to the topological arrangement of the conduction path of the three-dimensional vector field to form a spatiotemporal coupling feature set; The thermal-electric joint anomaly coefficient generation component is used to perform noise stripping on the spatiotemporal coupled feature set guided by the vector field, and to separate the random thermal gradient caused by environmental temperature changes by utilizing the directional consistency characteristics of the three-dimensional vector field. The purified feature set is processed by bidirectional convolution projection, which calculates the convolution of the temperature rise rate and the high-frequency component of the current in the time domain, and calculates the projection components of the heat flow direction and the current ripple envelope in the direction of the three-dimensional vector field in the spatial domain. The bidirectional output is processed by path-weighted integration, and the spatiotemporal correlation strength is fused according to the conduction path weight of the three-dimensional vector field to generate the original anomaly coefficient. The original anomaly coefficient is then processed by vector field baseline calibration, and the output value is dynamically corrected according to the normal historical path data covered by the three-dimensional vector field to generate the thermal-electric joint anomaly coefficient.

[0035] The working principle and beneficial effects of the above technical solution are as follows: The three-dimensional thermal topology vector generation component of this embodiment is used to process the three-dimensional vector field data obtained by infrared thermal imaging scanning through vector-temperature field conversion, mapping the direction attribute and modulus intensity of the scanning vector to a spatial temperature gradient distribution; the spatial temperature gradient distribution is processed by heat flow path verification, using the heat conduction direction preset in the three-dimensional vector field to constrain the actual temperature gradient direction, generating a purified temperature field that removes non-correlated thermal noise; the purified temperature field is processed by dynamic thermal topology modeling, extracting the temperature rise rate and heat flow direction of key nodes along the predefined conduction path of the three-dimensional vector field, generating a three-dimensional thermal topology vector; the spatiotemporal coupling feature set generation component is used to process the heat flow direction data of the three-dimensional thermal topology vector through current phase coupling, aligning the current ripple oscillation period with the heat flow direction change period; the aligned current features are processed by ripple-heat flow correlation extraction, and the spatial coordinates of the three-dimensional thermal topology vector are... The system binds the high-frequency amplitude and spectral entropy of the current at the corresponding time. The bound data undergoes spatiotemporal feature matrix reconstruction processing, and the current-thermal field parameters are arranged topologically according to the conduction path of the three-dimensional vector field to form a spatiotemporal coupled feature set. The thermal-electric joint anomaly coefficient generation component is used to perform noise stripping processing on the spatiotemporal coupled feature set guided by the vector field, and uses the directional consistency characteristics of the three-dimensional vector field to separate the random thermal gradient caused by environmental temperature changes. The purified feature set undergoes bidirectional convolution projection processing, calculating the convolution of the temperature rise rate and the high-frequency current component in the time domain, and calculating the projection components of the heat flow direction and the current ripple envelope in the direction of the three-dimensional vector field in the spatial domain. The bidirectional output undergoes path-weighted integral processing, and the spatiotemporal correlation strength is fused according to the conduction path weight of the three-dimensional vector field to generate the original anomaly coefficient. The original anomaly coefficient undergoes vector field baseline calibration processing, and the output value is dynamically corrected according to the normal historical path data covered by the three-dimensional vector field to generate the thermal-electric joint anomaly coefficient. The above scheme achieves joint analysis and anomaly detection of thermal and electrical features through multi-level data processing. The three-dimensional thermal topology vector generation component extracts physically plausible key heat conduction features by converting vector fields and temperature fields and constraining heat flow paths, eliminating noise interference and ensuring that subsequent analysis is based on the actual heat diffusion trend. The spatiotemporal coupling feature set generation component dynamically aligns current ripple features with heat flow direction and binds high-frequency current characteristics to three-dimensional spatial coordinates, forming a spatiotemporally correlated joint feature matrix, providing a data foundation for cross-domain analysis. The thermal-electrical joint anomaly coefficient generation component uses path constraints of the three-dimensional vector field to remove random noise and calculates the thermal-electrical coupling strength through temporal convolution and spatial projection. Finally, it dynamically calibrates the anomaly detection threshold by combining historical data, improving the robustness of detection.

[0036] In summary, this embodiment achieves joint anomaly detection of thermal and electric fields through thermal topology modeling, current feature coupling, and bidirectional anomaly analysis. Simultaneously, it utilizes the physical constraints of a three-dimensional vector field to optimize feature extraction and noise suppression, ensuring both spatial accuracy and temporal sensitivity in the detection results. This achieves deep coupling between predicted conduction paths and measured thermal field data. The three-dimensional vector field not only defines the scanning path but also serves as the physical link for thermal-electric feature fusion: its directional properties guide the generation of thermal topology vectors, the path topology constrains the spatiotemporal alignment dimension, and the fusion of conduction intensity weight anomaly coefficients forms a closed-loop logic of predicted path, measured verification, feature binding, and noise filtering.

[0037] Example 6: Based on Example 5, the thermo-electric combined anomaly coefficient generation component provided in this embodiment of the invention includes: The current-temperature rise sequence generation sub-component is used to purify the temperature rise rate data in the spatiotemporal coupling feature set after path propagation delay compensation processing. Based on the length and thermal conductivity of each conduction path in the three-dimensional vector field, the temperature rise time sequence of different spatial nodes is calibrated. The calibrated temperature rise sequence is then processed by high-frequency current component recombination, which rearranges the high-frequency components of the same source current according to the calibration time sequence of the conduction path to generate a completely synchronized current-temperature rise time domain sequence pair. The temporal convolution generation sub-component is used to process the current-temperature rise time-domain sequence pair through oscillation phase slicing, dividing it into independent oscillation period segments with the zero-crossing point of the high-frequency current component as the boundary; each oscillation period segment is processed by incremental energy accumulation to obtain the integral of the product of the temperature rise rate change and the high-frequency current amplitude change within a single period; the accumulation result is processed by multi-period sliding aggregation, and the incremental energy of three consecutive oscillation periods is slidably superimposed along the time axis to generate temporal convolution; The three-dimensional projection component generation sub-component is used to purify the heat flow direction data in the feature set. After vector field direction conformity processing, the cosine value of the angle between the measured heat flow direction and the preset direction of the three-dimensional vector field is compared. The cosine value of the angle is processed by current envelope spatial decomposition, and the amplitude distribution of the current ripple envelope is projected onto the main heat conduction direction defined by the three-dimensional vector field. The projection amplitude is processed by direction weight modulation and multiplied by the vector field direction conformity coefficient to generate the spatial projection component.

[0038] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the current-temperature rise sequence generation sub-component is used to purify the temperature rise rate data in the spatiotemporal coupling feature set after path propagation delay compensation processing. Based on the length and thermal conductivity of each conduction path in the three-dimensional vector field, the temperature rise sequence of different spatial nodes is calibrated. The calibrated temperature rise sequence undergoes high-frequency current component recombination processing, rearranging the high-frequency components of the same current source according to the calibration sequence of the conduction path to generate a completely synchronized current-temperature rise time-domain sequence pair. The time-domain convolution generation sub-component is used to process the current-temperature rise time-domain sequence pair through oscillation phase slicing, dividing it into independent oscillation period segments with the zero-crossing point of the current high-frequency component as the boundary. Each oscillation period segment is then amplified... The energy accumulation process yields the integral of the product of the temperature rise rate change and the high-frequency amplitude change of the current within a single cycle. The accumulated result undergoes multi-cycle sliding aggregation, where the incremental energy of three consecutive oscillation cycles is superimposed along the time axis to generate a time-domain convolution. A three-dimensional projection component generation sub-component is used to purify the heat flow direction data in the feature set. This data undergoes vector field direction conformity processing, comparing the cosine of the angle between the measured heat flow direction and the preset direction of the three-dimensional vector field. The cosine of the angle is then processed by current envelope spatial decomposition, projecting the amplitude distribution of the current ripple envelope onto the principal direction of heat conduction defined by the three-dimensional vector field. The projected amplitude is then processed by direction weight modulation and multiplied by the vector field direction conformity coefficient to generate a spatial projection component. These components achieve refined coupling analysis of thermal-electrical features through time-series calibration, time-domain integration, and spatial projection. The synergistic effects of each module are as follows: The current-temperature rise sequence generation sub-component eliminates temporal misalignment caused by spatial thermal diffusion differences through delay compensation of the heat conduction path and time-series reorganization of the current components, ensuring strict alignment of thermal-electrical data in the time dimension and providing synchronous input for subsequent correlation analysis. The temporal convolutional generation subcomponent calculates the energy coupling strength between temperature rise and current amplitude based on segmented current oscillation periods. It aggregates multi-period incremental energy through a sliding window to extract the temporal dynamic correlation features of the heat-electric interaction. The three-dimensional projection component generation subcomponent quantifies the spatial consistency of heat flow and current envelopes using vector field direction constraints. It highlights anomalous signals on the main conduction path and suppresses interference from irrelevant directions through direction weight modulation.

[0039] In summary, this embodiment achieves multi-dimensional joint anomaly feature extraction of thermal and electric fields through thermal-electric timing synchronization, temporal energy accumulation, and spatial orientation optimization. This enhances the adaptability of the detection results to the physical characteristics of the conduction path, while balancing temporal sensitivity and spatial resolution. Deep coupling is achieved through physically constrained dual-channel processing: the temporal channel quantifies the correlation strength between current disturbance and temperature rise into temporal convolution through three levels of processing: conduction path delay compensation, oscillation phase slicing, and incremental energy accumulation. Path propagation delay compensation ensures that the thermal inertia differences of different spatial nodes are accurately canceled. The spatial channel quantifies the matching metric between the spatial distribution of current and the direction of heat flow into projection components through three levels of processing: orientation conformity verification, envelope direction projection, and weight modulation. The preset direction of the three-dimensional vector field serves as the projection reference axis, and the orientation conformity coefficient serves as a physical consistency check. The final bidirectional output strictly follows the physical laws of the conduction path; temporal convolution characterizes energy transfer efficiency, and the spatial projection component characterizes spatial orientation consistency.

[0040] Example 7: Based on Example 6, the temporal convolution generator sub-component provided in this embodiment of the invention includes: The weighted energy sequence generation module is used to process the single-cycle energy value output by incremental energy accumulation processing. After the heat capacity mapping of the conduction path is processed, energy weight coefficients are assigned according to the heat capacity parameters of each path in the three-dimensional vector field. The energy weight coefficients are then processed by periodic energy modulation, multiplying the single-cycle incremental energy by the corresponding path heat capacity weight to generate a heat capacity-weighted periodic energy value. The coupled energy matrix generation module is used to calculate the heat conduction attenuation factor of the current period energy and the energy of the two preceding periods after the heat capacity-weighted period energy value is processed by the adjacent period conduction correlation processing. The heat conduction attenuation factor is then processed by energy coupling superposition processing to merge the current period weighted energy and the preceding period energy according to the attenuation factor ratio to generate a cross-period coupled energy unit. The cross-period coupled energy unit is then processed by time axis sliding window processing to generate a coupled energy matrix with three consecutive oscillation periods as the window unit. The aggregation result conversion module is used to calculate the path thermal interference effect of the coupled energy matrix. It calculates the energy interference coefficient based on the minimum spacing between adjacent conduction paths in the three-dimensional vector field. The energy interference coefficient is processed by matrix energy aggregation to perform interference compensation superposition on the cross-period energy of each path in the coupled energy matrix. The aggregation result is processed by time-domain energy density conversion, and the total energy value is divided by the time window width to generate a time-domain convolution.

[0041] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the weighted energy sequence generation module outputs a single-cycle energy value through incremental energy accumulation processing. This value undergoes conduction path thermal capacity mapping processing, and energy weight coefficients are assigned according to the thermal capacity parameters of each path in the three-dimensional vector field. These energy weight coefficients are then processed by periodic energy modulation, multiplying the single-cycle incremental energy by the corresponding path thermal capacity weight to generate a thermal capacity-weighted periodic energy value. The coupled energy matrix generation module processes the thermal capacity-weighted periodic energy value through adjacent-cycle conduction correlation processing, calculating the thermal conduction attenuation factor between the current periodic energy and the energy of the two preceding periods. The thermal conduction attenuation factor is then processed by energy coupling superposition, transforming... The current period weighted energy and the forward period energy are fused according to the attenuation factor ratio to generate a cross-period coupled energy unit. This cross-period coupled energy unit is processed through a time-axis sliding window, generating a coupled energy matrix with three consecutive oscillation periods as the window unit. The aggregation result conversion module is used to calculate the path thermal interference effect of the coupled energy matrix, calculating the energy interference coefficient based on the minimum distance between adjacent conduction paths in the three-dimensional vector field. The energy interference coefficient is then processed through matrix energy aggregation, performing interference compensation superposition on the cross-period energy of each path within the coupled energy matrix. The aggregation result is then processed through time-domain energy density conversion, dividing the total energy value by the time window width to generate a time-domain convolution. This scheme achieves refined modeling of thermal-electric energy in the time and spatial domains through thermal capacity weight allocation, cross-period energy coupling, and interference compensation. The weighted energy sequence generation module dynamically adjusts the single-period energy weight according to the thermal capacity parameters of the conduction path, highlighting the contribution of high thermal capacity paths and suppressing noise interference from low thermal capacity paths, ensuring that the energy characterization matches the physical conduction characteristics. The coupled energy matrix generation module correlates adjacent periodic energies through a thermal conduction attenuation factor, quantifies the diffusion and attenuation effect of energy over time, and integrates multi-period coupled energies using a sliding window to capture the temporal dynamic evolution of thermal-electric interactions. The aggregation result conversion module introduces an interference coefficient related to path spacing to compensate for the energy superposition or cancellation effect of adjacent conduction paths. Finally, through temporal energy density conversion, the aggregated energy is normalized to the equivalent intensity per unit time, eliminating the influence of the time window width.

[0042] In summary, this embodiment generates a convolution result that reflects both the thermal characteristics of the conduction path and the temporal energy propagation law through thermal capacity weight modulation, cross-period coupling attenuation, and spatial interference compensation, providing high-resolution temporal dynamic features for joint thermal-electric anomaly detection. This process achieves physically interpretable generation of the temporal convolution through third-order energy integration under thermophysical constraints: thermal capacity weight allocation, cross-period coupling, and path interference integration. The thermal capacity weight reflects the differences in energy accumulation characteristics of different conduction paths, the thermal conduction attenuation factor models the natural dissipation of energy over time, and the path interference coefficient quantifies the mutual influence of adjacent heat flow paths. The final output temporal convolution is essentially the anomalous energy density per unit time after path characteristic calibration, cross-period conduction correction, and spatial interference compensation. Its value simultaneously reflects the efficiency of current disturbance energy conversion into heat energy and the path conduction characteristics, forming the physical basis of the anomaly coefficient together with the spatial projection component.

[0043] Example 8: As Figure 5 As shown, based on Embodiment 1, the scheduling interaction subsystem provided in this embodiment of the invention includes: The anomaly coefficient processing component is used to process the combined thermal-electric anomaly coefficient through the heat capacity mapping of the conduction path, and to call the preset path heat capacity parameters in the three-dimensional vector field; the path heat capacity parameters are processed by the anomaly coefficient weighting, and the original anomaly coefficient is multiplied by the reciprocal of the path heat capacity to generate a heat capacity weighted anomaly coefficient that suppresses the interference of high heat capacity paths. The pattern matching processing component is used to compare the heat capacity weighted anomaly coefficient with historical failure mode matching processing and the historical anomaly coefficient and fault record of the same conduction path. The matching result is processed by the penetration intensity calculation, and a single path penetration intensity value is generated according to the standard deviation multiple of the current anomaly coefficient from the historical benchmark value. The single path penetration intensity value is processed by the path length attenuation, and the penetration intensity is attenuated exponentially according to the path length in the three-dimensional vector field to output the single path permeability.

[0044] The indicator output component is used to map the single-path permeability to three-dimensional spatial coordinates through topological expansion of the transmission path; the spatial permeability distribution is processed by risk field gradient calculation to extract the gradient extreme points of permeability spatial variation; the gradient extreme points are processed by neighborhood risk superposition to aggregate the permeability weighted sum of all paths within the radius centered on the extreme points; the aggregated results are processed by permeability intensity classification, and output three levels of risk permeability indicators (high, medium, and low) according to the preset physical range of the superposition value.

[0045] The spatial permeability distribution is generated by the topology unfolding of the conduction path: the single-path permeability (from the pattern matching processing component) is processed by path coordinate mapping, and the linear path permeability value is marked to the corresponding three-dimensional position based on the spatial coordinate data of the three-dimensional thermal topology vector; the marked data is processed by spatial interpolation, and permeability gradient transition values ​​are generated between adjacent paths according to the conduction path spacing defined by the three-dimensional vector field, and finally the three-dimensional spatial permeability distribution covering the scanned area is formed.

[0046] The working principle and beneficial effects of the above technical solution are as follows: The anomaly coefficient processing component in this embodiment is used to process the thermal-electrical joint anomaly coefficient through conduction path thermal capacity mapping, and calls the preset path thermal capacity parameters in the three-dimensional vector field; the path thermal capacity parameters are processed by anomaly coefficient weighting, multiplying the original anomaly coefficient by the reciprocal of the path thermal capacity to generate a thermal capacity weighted anomaly coefficient that suppresses interference from high thermal capacity paths; the mode matching processing component is used to process the thermal capacity weighted anomaly coefficient through historical failure mode matching, comparing the historical anomaly coefficients and fault records of the same conduction path; the matching result is processed by penetration intensity calculation, and a single-path penetration intensity value is generated based on the standard deviation multiple of the current anomaly coefficient from the historical benchmark value; the single-path penetration intensity value is processed by path length attenuation, and the penetration intensity is attenuated exponentially according to the path length in the three-dimensional vector field to output the single-path penetration. The index output component processes single-path permeability through topological expansion of the conduction path, mapping linear path permeability to three-dimensional spatial coordinates. Spatial permeability distribution is processed through risk field gradient calculation, extracting gradient extrema points of spatial permeability variation. These gradient extrema points are then processed through neighborhood risk superposition, aggregating the weighted sum of permeability across all paths within a radius centered on the extrema point. The aggregated result undergoes permeability intensity grading, outputting high, medium, and low-risk permeability indices based on the preset physical range of the superposition value. This multi-level processing achieves quantitative assessment and spatial positioning of abnormal risks in the thermal-electric combined system, constructing a complete risk analysis system from three dimensions: the anomaly coefficient processing component, through heat capacity parameter weighting, effectively suppresses the masking effect of high heat capacity paths on abnormal signals, highlighting true fault characteristics; the reciprocal heat capacity weighting mechanism makes conduction paths with different physical characteristics comparable, establishing standardized input for subsequent analysis. The pattern matching component establishes the correlation between anomaly coefficients and fault types by comparing historical data, and the penetration intensity calculation quantifies the degree to which the current anomaly deviates from normal operating conditions. Path length attenuation processing restores the physical attenuation law of fault signals during three-dimensional spatial transmission, ensuring that penetration reflects the true source risk. Topology unfolding processing reconstructs linear path data into a three-dimensional risk field, and gradient calculation accurately locates the core risk area. Neighborhood overlay processing overcomes the limitations of single-point assessment, reflecting the radiation range of fault impact through penetration aggregation. The three-level hierarchical output ultimately transforms complex physical quantities into intuitive risk decision-making basis.

[0047] In summary, this embodiment forms a progressive processing chain of signal optimization, pattern recognition, and spatial positioning, realizing a structured transformation from the original anomaly coefficient to a three-dimensional risk level. This provides a fault early warning framework for thermal-electric systems that balances physical accuracy and engineering practicality. This embodiment achieves precise risk index classification: heat capacity normalization eliminates the interference of material thermal inertia on the anomaly coefficient; path penetration modeling combined with historical fault data quantifies the risk transmission intensity; and spatial risk field superposition reveals the diffusion trend of abnormal heat sources in three-dimensional space. The final output risk penetration index is essentially a three-dimensional risk diffusion intensity value after heat capacity calibration, historical verification, and spatial aggregation. Its classification logic is based on the physical range of the spatially aggregated penetration value: high penetration corresponds to an aggregation value exceeding the material's tolerance limit, medium penetration corresponds to an aggregation value within the material's fatigue threshold range, and low penetration corresponds to an aggregation value below the safe operating baseline. This provides a classified physical basis for the bidirectional communication of the scheduling system.

[0048] Example 9: Based on Example 8, the indicator output component provided in this embodiment of the invention includes: The integral processing sub-component is used to process the spatial permeability distribution data after heat flow direction constraint processing, and to use the heat flow direction data of the three-dimensional thermal topology vector to limit the principal axis of gradient calculation; the permeability distribution after direction constraint is processed by anisotropic differentiation, and the spatial change rate of permeability is calculated along the principal direction of heat flow to generate a heat conduction gradient field; the heat conduction gradient field is processed by potential field integral transformation, and the gradient field is inversely integrated along the heat flow path to generate a risk potential field characterizing the intensity of risk accumulation; The verification processing sub-component is used to construct a three-dimensional continuous potential energy distribution surface by processing the potential energy field data through potential energy surface modeling; the potential energy distribution surface is processed by saddle point detection to identify critical points on the surface where the first derivative along the heat flow direction is zero and the second derivative is negative; the critical points are processed by heat flow path verification to filter points located on the preset conduction path of the three-dimensional vector field and output the coordinates of the gradient extremum points. The aggregation processing sub-component maps the gradient extreme point coordinates to the corresponding physical conduction paths in the 3D vector field after the conduction path association processing. The associated physical conduction paths are processed by the neighborhood thermal resistance network construction processing, connecting all physical conduction paths with heat flow within the radius centered on the polar coordinate point. The neighborhood thermal resistance network is processed by path thermal resistance weighting processing, calculating the weighting coefficient of each path according to the thermal resistance parameters of the material library. The weighting coefficients are processed by permeability aggregation processing, which performs thermal resistance weighted summation on the single-path permeability of all paths in the neighborhood to generate a spatial aggregated permeability value.

[0049] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the integral processing subcomponent is used to constrain the spatial permeability distribution data by heat flow direction, using the heat flow direction data of the three-dimensional thermal topology vector to limit the principal axis of gradient calculation; the permeability distribution after direction constraint undergoes anisotropic differential processing, calculating the spatial rate of change of permeability along the principal direction of heat flow to generate a heat conduction gradient field; the heat conduction gradient field undergoes potential field integral transformation processing, integrating the gradient field inversely along the heat flow path to generate a risk potential field characterizing the intensity of risk accumulation; the verification processing subcomponent is used to model the risk potential field data using potential energy surfaces, constructing a three-dimensional continuous potential energy distribution surface; the potential energy distribution surface undergoes saddle point detection processing to identify the points along the heat flow direction on the surface. The critical point is defined as having a first derivative of zero and a negative second derivative. After heat flow path verification, points located on pre-defined conduction paths in the 3D vector field are selected, and the coordinates of the gradient extremum points are output. An aggregation sub-component maps these gradient extremum point coordinates to the corresponding physical conduction paths in the 3D vector field after conduction path association processing. The associated physical conduction paths are then processed by a neighborhood thermal resistance network, connecting all heat flow-connected physical conduction paths within a radius centered on the polar coordinate point. The neighborhood thermal resistance network undergoes path thermal resistance weighting processing, calculating the weighting coefficients for each path based on the thermal resistance parameters in the material library. These weighting coefficients are then processed by permeability aggregation, summing the single-path permeability of all paths within the neighborhood using thermal resistance weighting to generate a spatial aggregated permeability value. This scheme first establishes a risk quantification foundation through integral processing. After axial constraints and gradient calculation, the spatial permeability data is transformed into an integrable conduction gradient field, ultimately generating a 3D potential field distribution reflecting the cumulative risk intensity, thus achieving a quantitative conversion from raw permeability to risk intensity. Secondly, key point location is completed through verification processing. Critical points conforming to thermodynamic characteristics are accurately identified on a three-dimensional potential energy surface, and their physical rationality is verified through pre-defined conduction paths. Finally, gradient extremum coordinates with practical conduction significance are output, ensuring consistency between the mathematical characteristics of key points and the physical scenario. Finally, regional permeability integration is achieved through aggregation processing, mapping discrete extremum coordinates back to the actual conduction path network. Through thermal resistance weighted calculation, the permeability of a single path is integrated into an aggregated value reflecting the overall permeability characteristics of the region; thus establishing a scale transformation from discrete points to a continuous field.

[0050] In summary, the cascading use of the three sub-components in this embodiment fully realizes the spatial transformation chain from raw permeability data to regional risk indicators: first, a risk potential field foundation is constructed; then, key transmission nodes are precisely located; and finally, regional-level permeability assessment indicators are integrated. The entire process maintains the spatial coherence of thermodynamic conduction characteristics, and the output aggregated permeability value can be directly used for risk assessment decisions.

[0051] Example 10: Based on Example 9, the aggregation processing sub-component provided in this embodiment of the invention includes: The dimensional integration module is used to process the path thermal resistance data in the neighborhood thermal resistance network through material conductivity mapping and call the thermal conductivity parameters of the corresponding conductors in the material library; the thermal conductivity parameters are processed through path conduction efficiency, and the thermal conductivity value is integrated with the path cross-sectional area and length to generate the path conduction efficiency coefficient. The proportion conversion module is used to process the path conduction efficiency coefficient after neighborhood heat flow balance and calculate the node heat flow conservation equation based on the connection topology of each path in the thermal resistance network; the node heat flow balance solution is processed by efficiency weight normalization and the conduction efficiency coefficient is converted into weight coefficient according to the proportion of the path in the neighborhood. The inertia compensation module is used to process the heat flow weight coefficient through permeability-heat flow coupling, multiplying the single-path permeability with the weight coefficient to characterize the contribution of that path to the risk propagation in the neighborhood; the coupled contribution is aggregated through heat flow direction processing, summing the contributions of all paths in the neighborhood along the main heat flow direction defined by the three-dimensional thermal topology vector; the summation result is processed through thermal inertia compensation, and the time accumulation effect of the aggregated value is corrected according to the specific heat capacity parameter of the path material to generate a spatial aggregated permeability value.

[0052] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the dimensional integration module is used to process the path thermal resistance data in the neighborhood thermal resistance network through material conductivity mapping, and call the thermal conductivity parameters of the corresponding conductors in the material library; the thermal conductivity parameters are processed through path conduction efficiency, and the thermal conductivity value is integrated with the path cross-sectional area and length through physical dimensions to generate the path conduction efficiency coefficient; the proportion conversion module is used to process the path conduction efficiency coefficient through neighborhood heat flow balance, and calculate the node heat flow conservation equation according to the connection topology of each path in the thermal resistance network; the node heat flow balance solution is processed through efficiency weight normalization, and the conduction efficiency coefficient is converted into a weight coefficient according to the proportion of the path in the neighborhood; the inertia compensation module is used to process the heat flow weight coefficient through permeability-heat flow coupling, and multiply the single path permeability with the weight coefficient to characterize the contribution of the path in the neighborhood risk propagation; the coupling contribution is processed through heat flow direction aggregation, and the contribution of all paths in the neighborhood is summed along the main heat flow direction defined by the three-dimensional thermal topology vector; the summation result is processed through thermal inertia compensation, and the time accumulation effect of the aggregation value is corrected according to the path material specific heat capacity parameter to generate a spatial aggregation permeability value. The above scheme transforms discrete path thermal resistance parameters into spatially correlated aggregate permeability values, and establishes a thermal resistance network analysis program that includes three dimensions: material properties, network topology, and dynamic propagation. This enables cross-scale calculations from microscopic path parameters to macroscopic thermal propagation characteristics, while simultaneously satisfying the requirements of thermodynamic conservation laws, network node equilibrium conditions, and correction for the time-cumulative effect of dynamic propagation.

[0053] Example 11: Based on Example 10, the proportion conversion module provided in this embodiment of the invention includes: The proportional allocation submodule is used to establish a physical balance at the intersection of thermal resistance network paths by processing the path conduction efficiency coefficient through heat flow conservation in the junction area, ensuring that the total heat flow input equals the total heat flow output. After the heat flow conservation balance is processed by conduction capacity allocation, the heat flow in the junction area is allocated according to the relative proportion of the conduction efficiency coefficient of each path, generating the heat flow ratio of each path. The convergence processing submodule is used to process the path heat flow ratio value through multi-level convergence coupling, and to transfer the heat flow distribution results of adjacent convergence zones according to the thermal resistance network topology hierarchy. The hierarchical data transfer is processed by heat flow direction consistency verification to ensure that the heat flow direction of each path is consistent with the main direction defined by the three-dimensional thermal topology vector. The verified heat flow ratio is processed by network balance convergence, and iteratively adjusted until the heat flow distribution of the entire network satisfies the conservation conditions of all convergence zones, generating the node heat flow balance solution. The normalization submodule is used to normalize the proportion of path heat flow in the node heat flow balance solution, converting the heat flow proportion of each path in the neighborhood into a value in the range of [0,1]. The normalized proportion is then processed by material thermal hysteresis compensation, and the influence of heat flow transfer delay is corrected according to the thermal diffusion rate parameter of the material library to generate heat flow weight coefficient.

[0054] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the proportional allocation submodule is used to establish a physical balance at the intersection of thermal resistance network paths by processing the path conduction efficiency coefficient through heat flow conservation in the junction area, ensuring that the total heat flow input equals the total heat flow output. This heat flow conservation balance is then processed by conduction capacity allocation, distributing the heat flow in the junction area according to the relative proportion of the conduction efficiency coefficients of each path, generating a heat flow percentage value for each path. The convergence processing submodule is used to process the path heat flow percentage value through multi-level junction coupling, transmitting the heat flow allocation results of adjacent junction areas according to the thermal resistance network topology hierarchy. The hierarchical data transmission is then processed by the heat flow... The process involves consistency verification to ensure that the heat flow direction of each path aligns with the principal direction defined by the 3D thermal topology vector. The verified heat flow percentage undergoes network equilibrium convergence processing, iteratively adjusted until the overall network heat flow distribution satisfies the conservation conditions of all intersection zones, generating a node heat flow equilibrium solution. A normalization submodule is used to normalize the path heat flow percentage values ​​in the node heat flow equilibrium solution, converting the heat flow percentage of each path within its neighborhood to a value in the [0,1] interval. The normalized percentage values ​​are then processed by material thermal hysteresis compensation, correcting for heat flow transfer delay based on the material library's thermal diffusion rate parameters, generating heat flow weight coefficients. This scheme transforms the path conduction efficiency coefficient into globally consistent heat flow weight coefficients, forming a heat flow distribution solution that simultaneously satisfies local node conservation, global network convergence, and directional consistency constraints. This process achieves the transition from local physical equilibrium to a network-wide steady-state solution, and ensures the physical rationality and computational compatibility of the weight coefficients through normalization and dynamic compensation.

[0055] Example 12: Based on Example 11, the proportional allocation submodule provided in this embodiment of the invention includes: The ratio processing unit is used to process the path conduction efficiency ratio in the heat flow conservation equilibrium to obtain the ratio of the conduction efficiency coefficient of each path in the intersection area to the maximum efficiency coefficient; the path conduction efficiency ratio is processed by the heat flow allocation weight generation process, and the square root of the ratio is mapped to the initial value of the allocation weight to generate the normalized allocation weight. The condition calibration unit is used to normalize the allocation weights. After the conservation condition calibration process, the weight values ​​are adjusted according to the total input heat flux in the junction area to ensure that the sum of the allocation weights of each path is equal to 1. The calibration weights are then processed by material interface effect correction, which calls the thermal reflection coefficients of different conductor connection interfaces in the material library to correct the heat flux loss at the interface. The corrected weights are then processed by heat flux ratio calculation, which multiplies the total input heat flux by the weight values ​​of each path to generate the physical allocation ratio. The density conversion unit is used to distribute the input heat flow to each output path according to the output path heat flow synthesis process. The distributed heat flow value is processed by path heat flow density conversion and divided by the cross-sectional area of ​​the path to obtain the heat flow per unit area. The heat flow per unit area is normalized by the proportion value to calculate the percentage of heat flow of each path to the total output heat flow of the junction area, and generate the path heat flow proportion value.

[0056] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the ratio processing unit processes the path conduction efficiency ratio to obtain the ratio of the conduction efficiency coefficient of each path in the junction area to the maximum efficiency coefficient, based on the heat flow conservation equilibrium. The path conduction efficiency ratio is then processed by heat flow allocation weight generation, mapping the square root of the ratio to the initial value of the allocation weight to generate a normalized allocation weight. The condition calibration unit processes the normalized allocation weight through conservation condition calibration, adjusting the weight value according to the total input heat flow in the junction area to ensure that the sum of the allocation weights for each path equals 1. The calibration weight is then processed by material interface effect correction, calling the material... The thermal reflection coefficient at the interface connecting different conductors in the mass library is used to correct for heat flow loss at the interface. The corrected weights are then processed by heat flow ratio calculation, multiplying the total input heat flow by the weight value of each path to generate a physical distribution ratio. The density conversion unit is used to process the output path heat flow synthesis and distribute the input heat flow to each output path proportionally. The distributed heat flow value is then processed by path heat flow density conversion and divided by the path cross-sectional area to obtain the heat flow per unit area. The heat flow per unit area is then normalized to calculate the percentage of heat flow in each path relative to the total output heat flow in the junction area, generating a path heat flow ratio value. The above scheme converts the conduction efficiency parameter into a heat flow distribution ratio with physical conservation and interface correction characteristics, and achieves three-dimensional spatial distribution quantification through heat flow density conversion.

[0057] Example 13: Based on Example 12, the condition calibration unit provided in this embodiment of the invention includes: The path allocation subunit is used to decompose the normalized weights into independent allocation factors corresponding to specific output paths after the corrected weight data has undergone path independence calculation. The independent allocation factors are then bound to the total heat flux to match the physical dimensions of each factor with the total input heat flux, generating dimensional path allocation factors. The conservation verification subunit is used to process the dimensional path allocation factor by multiplying it by the total heat flux input in the junction area to generate the original heat flux allocation value. The original heat flux allocation value is then processed by interface reflection loss compensation, which deducts the interface energy loss based on the interface thermal reflection coefficient of the material library. The compensated heat flux value is then processed by path heat flux conservation verification to ensure that the sum of all output path heat flux values ​​equals the total input heat flux, thus generating the path heat flux value. The constraint processing subunit is used to process the path heat flux value through the path heat flux density, divide it by the maximum allowable flow rate of the path to convert it into a dimensionless proportional value; the dimensionless proportional value is then processed by the proportional range constraint to linearly map the proportional value to the [0,1] physical interval to generate the physical allocation ratio.

[0058] The working principle and beneficial effects of the above technical solution are as follows: The path allocation subunit of this embodiment is used to process the corrected weight data through path independence calculation, decomposing the normalized weight into independent allocation factors corresponding to specific output paths; the independent allocation factors are bound by the total heat flux, so that the physical dimensions of each factor match the total input heat flux, generating a dimensional path allocation factor; the conservation verification subunit is used to process the dimensional path allocation factor through the total heat flux product, multiplying it by the total input heat flux of the intersection area to generate the original heat flux allocation value; the original heat flux allocation value is processed by interface reflection loss compensation, deducting interface energy loss according to the interface thermal reflection coefficient of the material library; the compensated heat flux value is processed by path heat flux conservation verification, ensuring that the sum of the heat flux values ​​of all output paths is equal to the total input heat flux, generating the path heat flux value; the constraint processing subunit is used to process the path heat flux value through path heat flux density, dividing it by the maximum allowable current carrying capacity of the path to convert it into a dimensionless proportional value; the dimensionless proportional value is processed by proportional range constraint, linearly mapping the proportional value to the [0,1] physical interval, generating the physical allocation ratio. The above scheme achieves refined allocation and physical constraint management of heat flux data. By solving for path independence, the composite weights are decomposed into physically meaningful independent allocation factors. The total heat flux is bound to ensure that each factor has actual engineering dimensions, realizing the conversion from abstract weights to physical parameters. When generating the original allocation value based on the product of the total heat flux, reflection loss compensation and conservation verification are performed simultaneously. This considers both energy loss caused by material interface characteristics and strictly adheres to the thermodynamic conservation law, ensuring the rationality of energy allocation at the physical level. The physical heat flux value is standardized into a dimensionless proportion through current-carrying capacity conversion, and then a physical allocation proportion within the range [0,1] is generated through interval mapping. This dual processing preserves the original heat flux characteristics while satisfying the boundary constraint requirements of engineering parameters.

[0059] In summary, this embodiment forms a closed-loop processing chain: starting with the physical decoupling of weight data, undergoing dynamic energy adjustment, and finally outputting a standardized allocation scheme that meets engineering constraints; it realizes the complete transformation of heat flow allocation from mathematical abstraction to engineering usability, while ensuring that the system meets the dual constraints of energy conservation and physical feasibility.

[0060] Example 14: As Figure 6 As shown, based on Examples 1-13, the power metering box security method based on multi-source data provided in this embodiment of the invention includes the following steps: S100: Continuously analyzes the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically defined. Only the physical nodes associated with the current abnormal circuit are captured with millisecond-level high-resolution thermal field. S200: Within the edge node, the data of the scanned area is deconstructed into a three-dimensional thermal topology vector in real time; the three-dimensional thermal topology vector is spatiotemporally aligned with the characteristics of the same source current ripple, and a thermal-electric joint anomaly coefficient is generated through a lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance, and automatically eliminate irrelevant interference such as ambient temperature fluctuations. S300: Based on the thermal-electrical joint anomaly coefficient, it autonomously generates a graded risk penetration index and drives bidirectional communication with the power grid dispatching system in real time: when the penetration is high, it actively pushes fault location maps; when the penetration is low, it only receives the operation strategy issued by the dispatching side and optimizes the trigger threshold of current ripple accordingly.

[0061] The working principle and beneficial effects of the above technical solution are as follows: This embodiment first continuously analyzes the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics automatically trigger and dynamically delineate the scanning area of ​​infrared thermal imaging, and only perform millisecond-level high-resolution thermal field capture on the physical nodes associated with the current abnormal circuit. Secondly, within the edge nodes, the data of the scanning area is deconstructed into a three-dimensional thermal topology vector in real time. The three-dimensional thermal topology vector is spatiotemporally aligned with the characteristics of the same source current ripple, and a thermal-electric joint anomaly coefficient is generated through a lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance, and automatically eliminate irrelevant interference such as ambient temperature fluctuations. Finally, based on the thermal-electric joint anomaly coefficient, a graded risk penetration index is autonomously generated to drive bidirectional communication with the power grid dispatching system in real time: when the penetration is high, the fault location map is actively pushed; when the penetration is low, only the operation strategy issued by the dispatching side is received, and the trigger threshold of the current ripple is optimized accordingly. The above scheme achieves accurate perception and intelligent decision-making regarding the safety status of power metering boxes through multi-source data collaborative processing and dynamic response mechanisms. First, by using spatiotemporal coupling detection of current ripple time-spectrum characteristics and infrared thermal imaging, a mapping relationship between electrical anomalies and physical thermal fields is established. A high-frequency oscillation mode triggering mechanism dynamically shrinks the thermal imaging scanning range to the fault-related area, reducing data processing load while ensuring detection accuracy. Second, thermal-electric joint analysis implemented at edge nodes constructs a spatiotemporal alignment model of three-dimensional thermal topology vectors and current characteristics. A lightweight coupling algorithm projects the heating location, temperature rise gradient, and current disturbance mode onto a unified dimension through feature space projection, forming a physically interpretable joint anomaly coefficient that effectively separates inherent equipment heating from environmental noise. Finally, a hierarchical communication strategy based on dynamic penetration indices enables adaptive interaction with the power grid dispatch system. Through nonlinear quantization of the anomaly coefficient, the system proactively uploads fault topology data under high-risk conditions and receives dispatch instructions and adjusts detection thresholds under normal operating conditions, achieving a balance between detection sensitivity and system resource utilization.

[0062] In summary, this embodiment achieves full-process protection from microscopic electrical feature capture to macroscopic system linkage, enhances fault identification specificity through spatiotemporal correlation analysis of multiple physical quantities, utilizes edge computing to achieve synergistic optimization of data dimensionality reduction and feature extraction, and constructs a hierarchical response system based on penetration index, ultimately achieving a balance between detection accuracy, computational efficiency, and system compatibility.

[0063] Example 15: As Figure 7 As shown, based on Embodiments 1-13, the power metering box safety device based on multi-source data provided in this embodiment of the invention includes: power metering box door 1, display screen 2, control box 3, and box body 4; Among them, the front left side of the box 4 is equipped with an energy metering box door 1, the right side of the energy metering box door 1 is equipped with multiple display screens 2, the side of the energy metering box door 1 is equipped with a control box 3, the control box 3 is equipped with an integrated circuit board, and the integrated circuit board stores the programs and instructions related to the current characteristic acquisition subsystem, the lightweight coupling subsystem and the scheduling interaction subsystem.

[0064] In this embodiment, the electricity metering box can be a single-phase nine-meter box, including two models: SX / PX-D901N and SX / PX-D902N. The first type of single-phase nine-meter box includes a molded case circuit breaker, disconnector, connector, and miniature circuit breaker. The molded case circuit breaker specification is 250L / 3300, the disconnector specification is 200A125 / 2P100AH, the connector specification is WJC-D, and the miniature circuit breaker specification is 63 / 2PC63A. The second type of single-phase nine-meter box includes a molded case circuit breaker, disconnector, and miniature circuit breaker. The molded case circuit breaker specification is 250L / 3300200A, the disconnector specification is 125 / 2P100A, and the miniature circuit breaker specification is 63 / 2PC63A. The electricity metering box can be a three-phase single-meter box, model SX / PX-S101XZ; it includes a disconnect switch, connectors, and miniature circuit breakers; the disconnect switch specification is 125 / 3P100A; the connector specification is HWJC-S; and the miniature circuit breaker specification is 80 / 4P80A. The electricity metering box can be a three-phase four-meter box, model SX / PX-S401N; it includes molded case circuit breakers, disconnect switches, and miniature circuit breakers; the molded case circuit breaker specifications are 250 / 3300250A; the disconnect switch specifications are 125 / 3P100A; and the miniature circuit breaker specifications are 80 / 4P80A.

[0065] The working principle and beneficial effects of the above technical solution are as follows: In this embodiment, the door 1 of the electricity metering box provides physical protection to ensure the safety of internal components; the display screen 2 displays the operating status and abnormal information of the electricity metering box in real time; the control box 3 has a built-in integrated circuit board that stores and executes program instructions for current characteristic acquisition, lightweight coupling, and scheduling interaction subsystems. The box body 4 serves as the supporting frame of the overall structure, integrating various components and ensuring stable operation.

[0066] In summary, the current characteristic acquisition subsystem of this embodiment continuously monitors current ripple and identifies abnormal high-frequency oscillation patterns; the lightweight coupling subsystem analyzes infrared thermal imaging data, calculates the thermal-electrical joint anomaly coefficient, and eliminates environmental interference; the scheduling interaction subsystem dynamically adjusts the communication strategy with the power grid dispatching system according to the anomaly level. Ultimately, this achieves real-time safety monitoring, accurate fault location, and intelligent dispatch response for the power metering box.

[0067] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of equivalents of this invention, this invention is also intended to include these modifications and variations.

Claims

1. A safety system for an electricity metering box based on multi-source data, characterized in that, Include: The lightweight coupling subsystem is used to deconstruct the data of the scanned area into a three-dimensional thermal topology vector in real time within the edge node; the three-dimensional thermal topology vector is spatiotemporally aligned with the ripple characteristics of the same source current, and a thermal-electric joint anomaly coefficient is generated through the lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance. The scheduling interaction subsystem is used to autonomously generate graded risk penetration indicators based on the thermal-electric joint anomaly coefficient, and drive bidirectional communication with the power grid scheduling system in real time: actively push fault location maps when the penetration is high. When the penetration rate is low, it only receives the operation strategy issued by the scheduling side and optimizes the trigger threshold of current ripple accordingly.

2. The power metering box safety system based on multi-source data as described in claim 1, characterized in that, The scheduling and interaction subsystem includes: The anomaly coefficient processing component is used to process the combined thermal-electric anomaly coefficient through the heat capacity mapping of the conduction path, and to call the preset path heat capacity parameters in the three-dimensional vector field; the path heat capacity parameters are processed by the anomaly coefficient weighting, and the original anomaly coefficient is multiplied by the reciprocal of the path heat capacity to generate a heat capacity weighted anomaly coefficient that suppresses the interference of high heat capacity paths. The pattern matching processing component is used to process the heat capacity weighted anomaly coefficient through historical failure mode matching. The matching results are then processed by permeability intensity calculation. Based on the standard deviation multiple of the current anomaly coefficient from the historical benchmark value, a single-path permeability intensity value is generated. The permeability intensity is then output based on the path length in the three-dimensional vector field, which decreases exponentially. The indicator output component is used to process the single-path permeability through the topological expansion of the transmission path, mapping the linear path permeability to three-dimensional spatial coordinates; extracting the gradient extreme points of permeability spatial variation; aggregating the weighted sum of permeability of all paths within the radius centered on the gradient extreme points; and outputting high, medium, and low risk permeability indicators according to the preset physical range of the superimposed value.

3. The power metering box safety system based on multi-source data as described in claim 2, characterized in that, The indicator output component includes: The integral processing subcomponent is used to process spatial permeability distribution data through heat flow direction constraints, and uses the heat flow direction data of the three-dimensional thermal topology vector to limit the principal axis of gradient calculation; and generates a heat conduction gradient field. The heat conduction gradient field is processed by potential field integration transformation. The gradient field is inversely integrated along the heat flow path to generate a risk potential field characterizing the intensity of risk accumulation. The verification processing sub-component is used to process the risk potential field data through potential energy surface modeling to construct a three-dimensional continuous potential energy distribution surface; the potential energy distribution surface is processed by saddle point detection to identify critical points on the surface; the critical points are processed by heat flow path verification to filter points located on the preset conduction path of the three-dimensional vector field and output the coordinates of the gradient extreme points. The aggregation processing sub-component maps the gradient extreme point coordinates to the corresponding physical conduction paths in the 3D vector field after the conduction path association processing. The associated physical conduction paths are processed by the neighborhood thermal resistance network construction processing, connecting all physical conduction paths with heat flow within the radius centered on the polar coordinate point. The neighborhood thermal resistance network is processed by path thermal resistance weighting processing, calculating the weighting coefficient of each path according to the thermal resistance parameters of the material library. The weighting coefficients are processed by permeability aggregation processing, which performs thermal resistance weighted summation on the single-path permeability of all paths in the neighborhood to generate a spatial aggregated permeability value.

4. The power metering box safety system based on multi-source data as described in claim 3, characterized in that, The aggregation processing sub-component includes: The dimensional integration module is used to process the path thermal resistance data in the neighborhood thermal resistance network through material conductivity mapping and call the thermal conductivity parameters of the corresponding conductors in the material library; the thermal conductivity parameters are processed through path conduction efficiency, and the thermal conductivity value is integrated with the path cross-sectional area and length to generate the path conduction efficiency coefficient. The proportion conversion module is used to process the path conduction efficiency coefficient after neighborhood heat flow balance and calculate the node heat flow conservation equation based on the connection topology of each path in the thermal resistance network; the node heat flow balance solution is processed by efficiency weight normalization and the conduction efficiency coefficient is converted into weight coefficient according to the proportion of the path in the neighborhood. The inertia compensation module is used to process the heat flow weight coefficient through permeability-heat flow coupling, multiplying the single-path permeability with the weight coefficient to characterize the contribution of that path to the risk propagation in the neighborhood; the coupled contribution is aggregated through heat flow direction processing, summing the contributions of all paths in the neighborhood along the main heat flow direction defined by the three-dimensional thermal topology vector; the summation result is processed through thermal inertia compensation, and the time accumulation effect of the aggregated value is corrected according to the specific heat capacity parameter of the path material to generate a spatial aggregated permeability value.

5. The power metering box safety system based on multi-source data as described in claim 4, characterized in that, The percentage conversion module includes: The proportional allocation submodule is used to establish a physical balance at the intersection of thermal resistance network paths by processing the path conduction efficiency coefficient through heat flow conservation in the junction area, ensuring that the total heat flow input equals the total heat flow output. After the heat flow conservation balance is processed by conduction capacity allocation, the heat flow in the junction area is allocated according to the relative proportion of the conduction efficiency coefficient of each path, generating the heat flow ratio of each path. The convergence processing submodule is used to process the path heat flow ratio value through multi-level convergence coupling and to transfer the heat flow distribution results of adjacent convergence areas according to the thermal resistance network topology hierarchy. The hierarchical data transmission undergoes heat flow direction consistency verification to ensure that the heat flow direction of each path is consistent with the main direction defined by the three-dimensional thermal topology vector. The verified heat flow ratio is then processed by network equilibrium convergence, and iteratively adjusted until the heat flow distribution of the entire network satisfies the conservation conditions of all intersection zones, generating a node heat flow equilibrium solution. The normalization submodule is used to normalize the proportion of path heat flow in the node heat flow balance solution, converting the heat flow proportion of each path in the neighborhood into a value in the range of [0,1]. The normalized proportion value is processed by material thermal hysteresis compensation, and the influence of heat flow transfer delay is corrected according to the thermal diffusion rate parameter of the material library to generate heat flow weight coefficient.

6. The power metering box safety system based on multi-source data as described in claim 5, characterized in that, The proportional allocation submodule includes: The ratio processing unit is used to process the path conduction efficiency ratio after heat flow conservation equilibrium, and obtain the ratio of the conduction efficiency coefficient of each path in the intersection area to the maximum efficiency coefficient. The path conduction efficiency ratio is processed by heat flux allocation weight generation, and the square root of the ratio is mapped to the initial value of the allocation weight to generate normalized allocation weights. The condition calibration unit is used to normalize the allocation weights after the conservation condition calibration process, and adjust the weight values ​​according to the total input heat flow in the junction area to ensure that the sum of the allocation weights of each path is equal to 1. The calibration weights are corrected for material interface effects by calling the thermal reflection coefficients of different conductor connection interfaces in the material library to correct for heat flow loss at the interface. The corrected weights are then processed by heat flow ratio calculation, multiplying the total input heat flow by the weight value of each path to generate the physical allocation ratio. The density conversion unit is used to physically distribute the input heat flow to each output path according to the output path heat flow synthesis process. The distributed heat flow value is processed by the path heat flow density conversion process and divided by the path cross-sectional area to obtain the heat flow per unit area. The heat flow per unit area is normalized by the percentage value, and the percentage of heat flow from each path to the total output heat flow in the junction area is calculated to generate the path heat flow percentage value.

7. The power metering box safety system based on multi-source data as described in claim 6, characterized in that, Condition calibration unit, comprising: The path allocation subunit is used to decompose the normalized weights into independent allocation factors corresponding to specific output paths after the corrected weight data has undergone path independence calculation. The independent allocation factors are bound to the total heat flux to match the physical dimensions of each factor with the total input heat flux, thus generating a path allocation factor with dimensions. The conservation verification subunit is used to process the dimensional path allocation factor by multiplying it by the total heat flux input in the junction area to generate the original heat flux allocation value. The original heat flux allocation value is then processed by interface reflection loss compensation, which deducts the interface energy loss based on the interface thermal reflection coefficient of the material library. The compensated heat flux value is then processed by path heat flux conservation verification to generate the path heat flux value. The constraint processing subunit is used to process the path heat flux value through the path heat flux density, divide it by the maximum allowable flow rate of the path to convert it into a dimensionless proportional value; the dimensionless proportional value is then processed by the proportional range constraint to linearly map the proportional value to the [0,1] physical interval to generate the physical allocation ratio.

8. The power metering box safety system based on multi-source data as described in claim 1, characterized in that, It also includes a current characteristic acquisition subsystem, which is used to continuously analyze the time-spectrum characteristics of the current ripple in the power metering box. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically defined, and only the physical nodes associated with the current abnormal circuit are captured with millisecond-level high-resolution thermal field.

9. A safety method for electricity metering boxes based on multi-source data, characterized in that, Includes the following steps: The time-spectrum characteristics of the current ripple inside the power metering box are continuously analyzed. When a specific high-frequency oscillation mode is detected, the time-spectrum characteristics are automatically triggered and the scanning area of ​​infrared thermal imaging is dynamically defined. Only the physical nodes associated with the current abnormal circuit are captured with millisecond-level high-resolution thermal field. Within the edge nodes, the data of the scanned area is deconstructed into a three-dimensional thermal topology vector in real time; the three-dimensional thermal topology vector is spatiotemporally aligned with the ripple characteristics of the same source current, and a thermal-electric joint anomaly coefficient is generated through a lightweight coupling algorithm to comprehensively characterize the correlation strength between the heating location, the temperature rise rate and the current disturbance. Based on the combined thermal and electrical anomaly coefficient, a graded risk penetration index is autonomously generated to drive real-time bidirectional communication with the power grid dispatch system: when the penetration is high, fault location maps are actively pushed. When the penetration rate is low, it only receives the operation strategy issued by the scheduling side and optimizes the trigger threshold of current ripple accordingly.

10. A safety device for an electricity metering box based on multi-source data, characterized in that, Includes: electricity metering box door, display screen, control box, and box body; The front left side of the enclosure has an electricity metering box door, the right side of the electricity metering box door has multiple displays, and a control box is installed on one side of the electricity metering box door. The control box contains an integrated circuit board, which stores programs and instructions related to the current characteristic acquisition subsystem, the lightweight coupling subsystem, and the scheduling interaction subsystem.

Images