Power distribution box fault online detection method and system based on internet of things, and medium

Abstract

The application discloses a power distribution box fault online detection method and system based on Internet of Things and a medium, and relates to the technical field of power distribution box fault detection. The method comprises the following steps: determining the conduction path of the current according to the actual connection structure of the internal conductor of the power distribution box and the on-off state of the switching device, dividing each conduction path into a plurality of continuous path segments; extracting the current propagation characteristics of each path segment; extracting the actual spatial trajectory of the current along each conduction path; mapping the multi-dimensional current propagation sequence to the spatial coordinates of the path segment, and comparing it with the actual spatial trajectory segment by segment to establish a consistency index; performing power distribution box fault matching according to the consistency index and the multi-dimensional current propagation sequence to establish an online detection result. The technical problem of insufficient fault positioning accuracy caused by the complex internal circuit structure of the power distribution box and the difficulty in timely and accurate identification of abnormal current state in the prior art is solved, and the technical effects of improving the accuracy and reliability of power distribution box fault online detection and positioning are achieved.

Description

Technical Field

[0001] This invention relates to the field of power distribution box fault detection technology, specifically to a method, system, and medium for online fault detection of power distribution boxes based on the Internet of Things. Background Technology

[0002] With the continuous improvement of the intelligence level of power systems, distribution boxes, as important equipment for power distribution and control, directly affect the safety and stability of the power supply system. Currently, fault detection in distribution boxes mainly relies on monitoring operating parameters such as current, voltage, and temperature, judging whether there are abnormalities in the equipment by detecting changes in these parameters. However, due to the complex internal wiring connections of distribution boxes, conductor distribution, switching device status, and load changes, fault characteristics can be diversified, making it difficult to accurately reflect internal abnormal states relying on a single detection parameter. Furthermore, some faults are not obvious in their initial stages, making it difficult for traditional detection methods to detect abnormal locations in a timely manner, easily leading to inaccurate fault location and detection lag. In addition, existing detection methods usually require manual inspection or power outage checks, which cannot meet the needs of long-term online monitoring and rapid fault diagnosis of distribution boxes. Summary of the Invention

[0003] This application provides an online fault detection method, system, and medium for distribution boxes based on the Internet of Things, which solves the technical problem in the prior art where the internal wiring structure of distribution boxes is complex and abnormal current states are difficult to identify in a timely and accurate manner, resulting in insufficient fault location accuracy.

[0004] The first aspect of this application provides a method for online fault detection of distribution boxes based on the Internet of Things, the method comprising: Based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, at least one current conduction path is determined. Each conduction path is divided into multiple continuous path segments according to physical connection nodes. At the nodes of each path segment, current time-series data is synchronously collected using non-invasive current sensors. Based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. The current propagation characteristics include at least time-domain waveform, frequency-domain characteristics, and power flow direction, establishing a multi-dimensional current propagation sequence for each conduction path. The feedback signals of the magnetic field sensor array deployed along the spatial distribution area of ​​the conduction path are read to establish a spatial magnetic field signal. The spatial partition of the current density is reconstructed using the magnetic field inverse source, and the actual spatial trajectory of the current along each conduction path is extracted. The multi-dimensional current propagation sequence is mapped to the spatial coordinates of the path segment and compared segment by segment with the actual spatial trajectory to establish a consistency index. Based on the consistency index and the multi-dimensional current propagation sequence, distribution box fault matching is performed to establish online detection results.

[0005] A second aspect of this application provides an online fault detection system for distribution boxes based on the Internet of Things, the system comprising: The system comprises the following modules: Path Division Module: Based on the actual connection structure of conductors inside the distribution box and the on / off state of switching devices, at least one current conduction path is determined, and each conduction path is divided into multiple continuous path segments according to physical connection nodes; Current Feature Extraction Module: At the nodes of each path segment, current time-series data is synchronously collected using non-invasive current sensors, and based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. The current propagation characteristics include at least time-domain waveform, frequency-domain characteristics, and power flow direction, establishing a multi-dimensional current propagation sequence for each conduction path; Current Trajectory Inversion Module: The feedback signal from the magnetic field sensor array deployed along the spatial distribution area of ​​the conduction path is read to establish a spatial magnetic field signal. The spatial partition of the inverse current density is reconstructed using the magnetic field inverse source, and the actual spatial trajectory of the current along each conduction path is extracted; Comparison Module: The multi-dimensional current propagation sequence is mapped to the spatial coordinates of the path segments and compared segment by segment with the actual spatial trajectory to establish a consistency index; Fault Matching Module: Based on the consistency index and the multi-dimensional current propagation sequence, fault matching of the distribution box is performed to establish online detection results.

[0006] A third aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the Internet of Things-based online fault detection method for distribution boxes provided in this application.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages: First, based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, at least one current conduction path is determined, and each conduction path is divided into multiple continuous path segments according to physical connection nodes. Next, non-invasive current sensors synchronously collect current time-series data at the nodes of each path segment. Based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. These current propagation characteristics include at least time-domain waveforms, frequency-domain characteristics, and power flow direction, establishing a multi-dimensional current propagation sequence for each conduction path. Then, feedback signals from magnetic field sensor arrays deployed along the spatial distribution area of ​​the conduction path are read to establish a spatial magnetic field signal. The spatial partitioning of the inverse current density is reconstructed using magnetic field inverse sources, and the actual spatial trajectory of the current along each conduction path is extracted. Further, the multi-dimensional current propagation sequence is mapped to the spatial coordinates of the path segments and compared segment by segment with the actual spatial trajectory to establish a consistency index. Finally, based on the consistency index and the multi-dimensional current propagation sequence, distribution box fault matching is performed to establish online detection results. This invention solves the technical problem in existing technologies where the complex internal wiring structure of distribution boxes makes it difficult to identify abnormal current states in a timely and accurate manner, resulting in insufficient fault location accuracy. It achieves the technical effect of improving the accuracy and reliability of online fault detection and location in distribution boxes. Attached Figure Description

[0008] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 A schematic flowchart of an online fault detection method for distribution boxes based on the Internet of Things provided in this application embodiment; Figure 2 This is a schematic diagram of the distribution path and current detection node distribution of the distribution box provided in the embodiments of this application; Figure 3 This is a schematic diagram illustrating the change in the current propagation sequence along the conduction path provided in an embodiment of this application. Figure 4 A schematic diagram of the structure of an IoT-based online fault detection system for distribution boxes provided in this application embodiment.

[0010] Figure labeling: 11 Conductive path division module, 12 Current feature extraction module, 13 Current trajectory inversion module, 14 Comparison module, 15 Fault matching module. Detailed Implementation

[0011] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0012] Example 1, as Figure 1 As shown, this application provides an online fault detection method for distribution boxes based on the Internet of Things, wherein the method includes: Based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, determine at least one conduction path for the current, and divide each conduction path into multiple continuous path segments according to the physical connection nodes.

[0013] In this embodiment, the actual connection relationships between conductors inside the distribution box and the real-time on / off status of each switching device are first obtained. The actual connection relationships include connection node information between the incoming terminal, busbar, branch conductors, terminals, and load connection terminals, as well as connection direction information between corresponding conductors. Based on the connection node information, a conductor topology structure inside the distribution box is constructed, with each conductor connection position as a topology node and the conductor connection segment between nodes as a topology edge. The topology structure is dynamically updated according to the on / off status of the switching devices. When a switching device is in the on state, the corresponding switching node is set as a valid connection node; when a switching device is in the off state, the corresponding connection edge is removed from the current valid topology structure. Further, the input terminal of the distribution box is used as the starting node of the current path, and each load output terminal or return terminal is used as the target node. In the updated valid... In the topology, a path search method is used to determine at least one conducting path through which the current actually passes, and the node sequence and connection order corresponding to each conducting path are recorded. When multiple branch loads are powered simultaneously at the same input terminal, parallel conducting paths are established according to the conductor connection relationship of each branch to characterize the current propagation relationship in different power supply branches. After obtaining each conducting path, the path is segmented according to the position of the physical connection node on the conducting path. The continuous conductor area between two adjacent physical connection nodes is divided into a path segment, and a unique identifier is established for each path segment. The unique identifier includes at least the starting node, ending node, spatial coordinate range, and the number of the conducting path to which it belongs, thereby forming a set of conducting paths composed of multiple continuous path segments, providing a path basis for subsequent current timing acquisition, current propagation feature extraction, and spatial trajectory consistency analysis.

[0014] At the nodes of each path segment, current timing data is synchronously collected using non-invasive current sensors. Based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. The current propagation characteristics include at least time-domain waveform, frequency-domain characteristics, and power flow direction, thus establishing a multi-dimensional current propagation sequence for each conduction path.

[0015] In this embodiment, based on the aforementioned determined conduction path and corresponding path segment division results, non-invasive current sensors are deployed at the start and end nodes of each path segment. These non-invasive current sensors acquire the current response signal generated around the corresponding conductor through electromagnetic induction and synchronously collect data from each node according to a unified sampling clock, obtaining current time-series data containing node identifiers, sampling time, current amplitude, and phase information. For adjacent nodes on the same conduction path, based on the spatial connection and time synchronization relationships between nodes, the current time-series data collected by the upstream node and the current time-series data collected by the downstream node are time-aligned, and the changes in current amplitude, phase delay, and energy transfer between adjacent nodes are calculated to form the current transfer relationship corresponding to the path segment. Further... The system performs feature analysis on the node current time-series data of each path segment, extracting current waveform change trends, peak value changes, periodic characteristics, and transient change characteristics from the time domain dimension. From the frequency domain dimension, it performs spectrum transformation on the current time-series data to extract fundamental components, harmonic distribution, high-frequency disturbance components, and frequency energy change characteristics. At the same time, it determines the actual current transmission direction based on the voltage reference direction, current phase relationship, and power calculation results between adjacent nodes, obtaining the power flow characteristics of the corresponding path segment. The time domain waveform characteristics, frequency domain characteristics, and power flow characteristics corresponding to each path segment are combined according to the path node sequence and associated with the spatial location, timestamp, and path number of the corresponding path segment to form a multidimensional current propagation sequence continuously distributed along the conduction path, which is used to characterize the propagation state of the current inside the distribution box in different path segments.

[0016] like Figure 2 , Figure 3 As shown, taking a conductive path inside a distribution box as an example, based on the actual connection structure of the conductors inside the distribution box and the on / off states of the switching devices QF1 and QF2, the current conductive path from the power supply side to the load side is determined. The conductive path is divided into multiple continuous path segments according to the physical connection nodes N1, N2, N3, and N4 on the conductive path. Non-intrusive current sensors I1, I2, I3, and I4 are set at the corresponding positions of each physical connection node. The current timing data at each node is synchronously acquired according to a unified sampling clock. The current timing data corresponding to adjacent nodes on the same conductive path are correlated to extract the current propagation characteristics of each path segment. Then, the path segments are continuously combined according to the arrangement order along the conductive path to establish the corresponding multidimensional current propagation sequence. Figure 3Taking the current amplitude dimension as an example, the propagation and change of current along multiple continuous path segments are shown. Under normal conditions, the current amplitude of each path segment changes continuously according to the normal transmission relationship. When some path segments are abnormal, the current amplitude of the corresponding path segment and its subsequent path segments undergoes abnormal attenuation or sudden change relative to the normal propagation state, thus forming a current propagation change characteristic that is different from the normal state.

[0017] The feedback signals of the magnetic field sensor array deployed in the spatial distribution area along the conduction path are read to establish the spatial magnetic field signal. The spatial partition of the inverse current density is reconstructed using the magnetic field inverse source, and the actual spatial trajectory of the current along each conduction path is extracted.

[0018] In this embodiment, a magnetic field sensor array is set along the spatial distribution area corresponding to each conduction path. The magnetic field sensor array is arranged around the conductor inside the distribution box at preset spatial intervals to collect magnetic field response signals generated by the conductor current at different spatial locations. Each magnetic field sensor is calibrated according to a unified spatial coordinate system, and a magnetic field feedback signal containing sensor spatial coordinates, sampling time, and magnetic field strength and direction information is obtained through synchronous sampling. Further, the magnetic field feedback signal undergoes time synchronization, spatial coordinate correction, and background magnetic field elimination processing to remove interference from the fixed magnetic field inside the distribution box, obtaining spatial magnetic field signals corresponding to the spatial locations of each conduction path. Based on the correspondence between the magnetic field generated by the conductor current and the current density distribution, a magnetic field response model is established. The spatial magnetic field signal is used as input, and the unknown current source distribution is solved by inverting the magnetic field inverse source reconstruction method. Multiple three-dimensional reconstruction units are divided according to the spatial region where the conduction path is located. The current density within each reconstruction unit is used as the variable to be solved. Based on the magnetic field strength, direction information collected by the magnetic field sensor array, and the spatial position relationship of the sensors, a magnetic field response moment is constructed. The current density values ​​of each spatial unit are calculated using an iterative optimization method to obtain the spatial distribution results of the current density inside the distribution box. Further, based on the spatial distribution results, the current density amplitudes in each spatial unit are filtered, and spatial connectivity analysis is performed on continuous high current density regions to extract the spatial connection paths of the current density peak regions. These paths are then combined with the spatial coordinate information of the aforementioned conduction paths to perform coordinate mapping, obtaining the spatial trajectory formed by the actual propagation of current along each conduction path. When there is a spatial offset between the actual current trajectory and the preset conductor connection path, the spatial coordinate information of the corresponding offset region is retained as the basis for subsequent current propagation consistency analysis and fault location judgment.

[0019] The multidimensional current propagation sequence is mapped to the spatial coordinates of the path segment and compared segment by segment with the actual spatial trajectory to establish a consistency index.

[0020] Furthermore, the consistency index includes at least one of trajectory offset, current distribution width variation coefficient, and transfer function distortion.

[0021] In this embodiment, firstly, the path number, node number, timestamp, and spatial coordinate information of each path segment in the multidimensional current propagation sequence are obtained. The time-domain waveform, frequency-domain characteristics, and power flow direction features of each path segment are then associated with their corresponding spatial locations, forming a current propagation feature distribution with spatial constraints. Further, based on the three-dimensional coordinate information of each trajectory point in the actual spatial trajectory, the current propagation feature distribution and the actual spatial trajectory are coordinate unified. The center position, lateral distribution range, and current propagation direction of the path segments in the multidimensional current propagation sequence are mapped to the coordinate system of the actual spatial trajectory, establishing a path segment-level current propagation spatial model. After completing the spatial mapping, according to the node order of the conducting path, the current propagation spatial model corresponding to the same numbered path segment is matched segment by segment with the actual spatial trajectory, calculating the current propagation region within each path segment and the actual spatial trajectory. The system considers the positional deviation, distribution range differences, and propagation state changes between different locations. Specifically, it obtains the trajectory offset, characterizing the degree of current propagation positional deviation, based on the spatial distance change between the current propagation center region and the actual spatial trajectory centerline. It calculates the change in current distribution width relative to the normal trajectory width based on the matching relationship between the lateral current diffusion range of each path segment and the actual conductor spatial boundary range, forming a current distribution width variation coefficient. It compares the current amplitude attenuation relationship, phase transmission relationship, and spatial continuity transmission relationship between adjacent path segments with the reference transmission relationship under normal conduction conditions to obtain the transfer function distortion degree. At least one of the trajectory offset, current distribution width variation coefficient, and transfer function distortion degree is combined to establish a comprehensive evaluation index characterizing the consistency between the current propagation state and the actual spatial trajectory, thereby achieving a quantitative judgment of abnormal conduction states inside the distribution box.

[0022] Furthermore, the trajectory offset is calculated based on a point-by-point comparison between the time-domain waveform and the actual spatial trajectory, including: The amplitude envelope and zero-crossing sequence of the current time-domain waveform of each path segment are extracted from the multidimensional current propagation sequence to establish the behavioral fingerprint of the current in the time domain for the corresponding path segment. Based on the amplitude envelope change rate and zero-crossing density distribution in the behavioral fingerprint, an equivalent current distribution function characterizing the intensity of the current distribution in the cross-section of the path segment is constructed, and the equivalent current distribution function is continuously expanded along the extension direction of the path segment to form a current spatial distribution expression. The current spatial distribution expression is weighted and integrated in the direction perpendicular to the conduction path, and the connected trajectory of the region with the maximum current density is extracted as the current centerline. Based on the spatial coordinates, the current centerline is aligned to the corresponding spatial position of the actual spatial trajectory according to the timestamp, and the offset distance sequence between the current centerline and the spatial trajectory centerline in the direction perpendicular to the path is calculated. The root mean square value of the offset distance sequence is output as the trajectory offset. When a non-periodic abrupt change occurs in the time domain waveform and the actual spatial trajectory does not change accordingly, it is determined that the current behavior is mismatched with the actual spatial trajectory distribution, and an unexpected conduction result is output.

[0023] Preferably, the current time-domain sampling data in the multidimensional current propagation sequence corresponding to each path segment is obtained, and the current sampling sequence of each path segment is represented as follows: The current sampling sequence is parsed to calculate the current amplitude envelope at the corresponding time point, thus obtaining the amplitude envelope sequence. The amplitude envelope sequence is used to describe the trend of current intensity change over time in the path segment; simultaneously, the current signs of adjacent sampling points are compared, and when adjacent sampling points satisfy... At that time, the corresponding sampling time is recorded as the zero-crossing point, forming a zero-crossing sequence. According to the amplitude envelope sequence The amplitude change rate between adjacent sampling times is calculated, and the zero-crossing density is calculated based on the ratio of the number of zero-crossings to the window length within a preset time window to obtain a behavioral fingerprint characterizing the current time-domain change state. The amplitude change rate and the zero-crossing density are used to describe the degree of current amplitude change and frequency change, respectively.

[0024] Based on the amplitude change rate and zero-crossing density in the behavioral fingerprint, the cross-section of each path segment is divided into multiple spatial sampling units, and an equivalent current distribution function is established according to the current response intensity corresponding to each spatial sampling unit. The equivalent current distribution function is expressed as follows: ,in, Indicates path location Cross-sectional coordinates The corresponding equivalent current distribution intensity, Indicates the rate of change of amplitude envelope. Let a represent the zero-crossing density, and let a and b represent the amplitude change characteristic weight and frequency change characteristic weight, respectively. The values ​​of a and b are 0.4 and 0.6, respectively, and a+b=1.

[0025] The equivalent current distribution function at each location is discretized and expanded along the conduction path. The cross-sectional current distribution results corresponding to different path locations are then stitched together in spatial coordinate order to generate a three-dimensional current spatial distribution matrix. Within each path segment cross-section, the maximum current density region is searched in the three-dimensional current spatial distribution matrix. Continuous high-response regions are extracted according to a preset current density filtering threshold, where the current density filtering threshold is set to 80% of the maximum current density value of the current cross-section. Spatial connectivity analysis is performed on the continuous high-response regions to obtain concentrated current density regions continuously distributed along the conduction path. The spatial center point of the concentrated current density region in each cross-section is calculated, and the spatial center points are connected according to the path direction to obtain the current centerline.

[0026] Furthermore, the current centerlines are sorted according to the acquisition timestamps and time-synchronized with the actual spatial trajectory obtained from the reconstruction of the magnetic field inverse source; for each current centerline sampling point... In the actual spatial trajectory, obtain trajectory points with the same timestamp. And determine the path direction vector based on the spatial extension direction of the conduction path at the current sampling point. ; Calculate the spatial difference vector between the sampling point of the current centerline and the actual spatial trajectory point. The spatial difference vector is along the path direction vector. Projective decomposition is performed to obtain the lateral offset component perpendicular to the conduction path direction. The calculation process is as follows: ,in, This represents the lateral offset distance between the current centerline at the i-th sampling point and the actual spatial trajectory. This represents the spatial coordinates of the current centerline at the i-th sampling point. This represents the spatial coordinates of the actual spatial trajectory at the corresponding timestamp. The first term represents the conduction path direction vector at the i-th sampling point, and the second term represents the projection component of the spatial difference vector along the conduction path direction.

[0027] The lateral offset distances corresponding to each sampling point are arranged sequentially to form an offset distance sequence. And calculate the root mean square value of the offset distance sequence to obtain the trajectory offset: ,in, The value represents the trajectory offset, and n represents the number of trajectory matching points involved in the calculation. The number of trajectory matching points is set to 500 to ensure that the trajectory offset calculation result can reflect the overall offset status of the path segment.

[0028] During the calculation of trajectory offset, the changes in the current time-domain waveform and the actual spatial trajectory are synchronously judged. When the rate of change of amplitude envelope within the continuous sampling window exceeds the preset amplitude change threshold, and the change in zero-crossing density exceeds the preset frequency change threshold, while the change in the position of the actual spatial trajectory within the corresponding time window is less than the preset trajectory change threshold, it is determined that the current propagation behavior of the current segment is inconsistent with the actual spatial trajectory distribution. An unexpected conduction result is output, and the corresponding path segment number and spatial coordinate position are recorded. The continuous sampling window time length is set to 50ms, the preset amplitude change threshold is set to twice the average value of the amplitude envelope change rate under normal operating conditions, the preset frequency change threshold is set to three times the average value of the zero-crossing density change under normal operating conditions, and the preset trajectory change threshold is set to 1mm. When the above conditions are met, it indicates that the current time-domain propagation state has undergone an abnormal change, while the actual spatial trajectory has not changed accordingly, thus confirming the existence of an unexpected conduction state.

[0029] Furthermore, the calculation of the coefficient of variation of the current distribution width includes: Based on the amplitude envelope in the multidimensional current propagation sequence, the lateral current diffusion range corresponding to each path segment is determined, and a current distribution width reference region is established according to the spatial boundary position of the corresponding path segment in the actual spatial trajectory. The lateral current diffusion range of each path segment is projected onto the spatial coordinate system where the actual spatial trajectory is located, and the overlap width and deviation width between the current distribution range and the current distribution width reference region are calculated based on the projection results to form a current distribution width sequence under spatial constraints. The current distribution width sequences of adjacent path segments are continuously compared along the conduction path direction, and the dispersion of the current distribution width of each path segment relative to the width of the actual spatial trajectory is calculated to construct the current distribution width variation coefficient.

[0030] Preferably, the current amplitude envelope sequence corresponding to each path segment in the multidimensional current propagation sequence is obtained. The amplitude envelope sequence is segmented according to a preset sampling time window to obtain the current amplitude distribution data corresponding to each time window. The cross-sectional area corresponding to each path segment is divided into multiple spatial sampling units. The current diffusion boundary is determined according to the amplitude response value corresponding to each spatial sampling unit. Among them, the spatial sampling units whose amplitude response value is greater than 40% of the maximum amplitude envelope value of the current path segment are determined as effective current distribution areas, and continuous spatial sampling units that meet this condition are also considered as effective current distribution areas. At the same time, the lateral current diffusion width is determined according to the maximum spacing of the effective current distribution areas in the cross-sectional direction. Simultaneously, based on the actual spatial trajectory data obtained from the magnetic field inverse source reconstruction, the coordinates of the spatial boundaries on both sides of the corresponding path segment are acquired, and the width of the actual conductor region in the cross-sectional direction is calculated to establish a reference region for the current distribution width of the corresponding path segment. .

[0031] Map the lateral current diffusion range corresponding to each path segment to the coordinate system of the actual spatial trajectory, based on the path segment spatial coordinate transformation matrix. The coordinates of the current distribution region are transformed so that the current diffusion region and the actual spatial trajectory are in the same spatial reference frame; the spatial overlap between the coordinate-transformed current distribution region and the current distribution width reference region is calculated to obtain the overlap width. Simultaneously, the deviation width of the current diffusion region extending beyond the boundary of the reference region is calculated. The actual current distribution width of the path segment is calculated based on the overlap width and the deviation width. ,in, This represents the actual current distribution width corresponding to the i-th path segment. This represents the reference value for the actual spatial trajectory width of the corresponding path segment. This indicates the overlap width between the current distribution region and the reference region. This indicates the width by which the current distribution area deviates from the reference area.

[0032] The actual current distribution widths corresponding to each path segment are sorted according to the path segment number along the conduction path direction to form a current distribution width sequence: ; Continuously compare the current distribution widths between adjacent path segments and calculate the dispersion of the current distribution width of each path segment relative to the overall path width trend; First, calculate the average value of the current distribution width sequence: Then calculate the standard deviation of the current distribution width: The coefficient of variation of the current distribution width is obtained based on the ratio between the standard deviation and the average width. ,in, This represents the coefficient of variation of the current distribution width, where n represents the number of path segments included in the statistics; when When the current distribution width variation threshold is exceeded, it is determined that the current spatial diffusion state on the current conduction path has changed abnormally. The number of path segments is determined based on the number of conductor connection nodes inside the distribution box, the effective current amplitude distribution threshold is set to 40% of the maximum amplitude envelope, and the width variation threshold is set to 0.3. When the current distribution width variation coefficient is greater than the width variation threshold, it indicates that the actual current diffusion range has changed significantly relative to the normal conductor space region, and the corresponding path segment is identified as an abnormal candidate region.

[0033] Furthermore, the calculation of the transfer function distortion includes: Based on the current timing data of adjacent path segments on the same conduction path, the amplitude transfer relationship and phase transfer relationship between adjacent nodes are obtained, and the current transfer characteristics corresponding to the path segment are established. According to the spatial continuity relationship between adjacent path segments in the actual spatial trajectory, the spatial transfer range corresponding to the current in the normal conduction state is determined, and a reference transfer characteristic matching the spatial transfer range is established. The current transfer characteristics and the reference transfer characteristics are compared on a path segment by path, and the amplitude attenuation deviation, phase delay deviation and spatial position offset during the actual current propagation process are extracted and fused to form the transfer function distortion degree.

[0034] Preferably, current time-series data collected from corresponding nodes of adjacent path segments on the same conduction path are obtained, wherein, for the i-th path segment, the current sequence of the upstream node is obtained. and downstream node current sequence The current time-series data are subjected to Fast Fourier Transform (FFT) to obtain the corresponding frequency domain response: , Based on the amplitude and phase relationships of the frequency domain responses of adjacent nodes, calculate the current transfer function between adjacent path segments: ,in, Let represent the current transfer function corresponding to the i-th path segment. The magnitude part of the transfer function represents the magnitude attenuation relationship of the current during the propagation along the path, and the phase part represents the phase delay relationship of the current during the propagation.

[0035] Furthermore, feature extraction is performed on the current transfer function to obtain amplitude transfer features. and phase transfer characteristics ,in: , Simultaneously, based on the actual spatial trajectory data obtained from the magnetic field inverse source reconstruction, the spatial connection direction, path length, and trajectory continuity relationship between adjacent path segments are acquired; based on the spatial connection relationship between adjacent path segments under normal conduction conditions, a corresponding reference spatial transmission range is established, including the reference path length. Reference dissemination direction and reference space offset range Based on the reference space transmission range and the current propagation data under historical normal operating conditions, establish reference transmission characteristics that match the current path segment. The reference transfer feature includes the reference amplitude transfer relationship. Reference phase transfer relationship and reference space propagation position .

[0036] The current propagation characteristics are compared with the reference propagation characteristics segment by segment to calculate the amplitude attenuation deviation, phase delay deviation, and spatial position offset during the actual current propagation process; wherein, the amplitude attenuation deviation is calculated as follows: Where N represents the number of frequency sampling points involved in the calculation. This represents the k-th frequency sampling point.

[0037] The phase delay deviation is calculated as follows: ,in, This represents the phase response of actual current propagation. This represents the reference phase response.

[0038] Spatial position offset based on the actual current propagation trajectory point Propagation position in reference space Distance calculation between: .

[0039] Furthermore, the amplitude attenuation deviation, phase delay deviation, and spatial position offset are fused according to preset weights to obtain the transfer function distortion degree corresponding to the path segment: ,in, This represents the distortion degree of the transfer function of the i-th path segment. , , These represent the amplitude deviation weight, phase deviation weight, and spatial offset weight, respectively. In the specific implementation process, , , They were set to 0.4, 0.35, and 0.25 respectively.

[0040] When the distortion of the transfer function When the distortion threshold is exceeded, the current propagation relationship of the corresponding path segment is determined to deviate from the normal state, and the path segment is identified as a candidate fault region; wherein, the preset distortion threshold is set to 0.3.

[0041] Based on the consistency index and multidimensional current propagation sequence, fault matching of distribution boxes is performed, and online detection results are established.

[0042] Furthermore, based on consistency indicators and multidimensional current propagation sequences, distribution box fault matching is performed to establish online detection results, including: When the trajectory offset and the coefficient of variation of the current distribution width both show significant anomalies, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence exhibits a non-periodic abrupt change and amplitude envelope attenuation, it is determined that there is an abnormal contact resistance fault in the path segment, and the spatial coordinates and severity level of the path segment are output.

[0043] Preferably, the trajectory offset corresponding to each path segment is obtained. Coefficient of variation of current distribution width and the time-domain waveform characteristics in the multidimensional current propagation sequence; the trajectory offset Offset threshold from preset trajectory Compare the coefficients of variation of the current distribution width. Compared with the preset width variation threshold Compare; when and When the current spatial propagation position and lateral diffusion range of the corresponding path segment are determined to have undergone abnormal changes, the trajectory offset threshold is... The width variation threshold is set to 2mm. Set it to 0.3.

[0044] Furthermore, the current time-domain waveform corresponding to the path segment is continuously sampled and analyzed to extract the amplitude envelope sequence. Calculate the amplitude change between adjacent sampling windows. And determine whether there is a non-periodic abrupt change in the current waveform; when the amplitude change within the continuous sampling window satisfies: ,in, This represents the average amplitude envelope value of the corresponding path segment under normal operating conditions; and the amplitude envelope sequence shows a continuous decay trend within three consecutive sampling windows, i.e.: When the trajectory offset and the current distribution width variation coefficient are both abnormal, it is determined that the current path segment has a current concentration area offset and an effective conductive cross-sectional area change.

[0045] Furthermore, fault type matching is performed based on the aforementioned spatial anomaly characteristics and current propagation attenuation characteristics; when the trajectory offset... Coefficient of variation of current distribution width If all values ​​exceed the corresponding threshold, and the time-domain waveform of the corresponding path segment in the multidimensional current propagation sequence exhibits a non-periodic abrupt change while the amplitude envelope continuously decays, then an abnormal contact resistance fault is determined to exist in that path segment. The fault location is determined based on the spatial coordinate range corresponding to the abnormal path segment, and the fault severity level is calculated based on the trajectory offset, the coefficient of variation of the current distribution width, and the degree of amplitude decay. The fault severity level is calculated as follows: ,in, This indicates the severity level of the abnormal contact resistance. Indicates the trajectory offset. Indicates the trajectory offset threshold. Indicates the coefficient of variation of the current distribution width. Indicates the width variation threshold. This represents the amplitude envelope attenuation. This represents the average amplitude envelope value under normal conditions. α, β, and γ represent the weights corresponding to trajectory offset, width change, and amplitude attenuation, respectively. In the specific implementation process, α=0.4, β=0.35, and γ=0.25 are set.

[0046] Based on the severity level evaluation value Classify fault levels, when At that time, it was determined to be a slight contact resistance abnormality; when At that time, it was determined to be a moderate contact resistance abnormality; when When the fault is detected, it is determined to be a severe contact resistance anomaly, and the corresponding path segment number, spatial coordinate range, and fault severity level are output, forming an online detection result for contact resistance anomaly.

[0047] Furthermore, the process of performing distribution box fault matching based on consistency indicators and multidimensional current propagation sequences, and establishing online detection results, also includes: When the distortion of the transfer function is significantly abnormal, and the frequency domain characteristics of the corresponding path segment in the multidimensional current propagation sequence show a sudden increase in the amplitude of high-frequency components or a drastic change in harmonic content, it is determined that there is a partial discharge or arc fault in the path segment, and the path segment where the discharge point is located is located based on the peak value of the distortion frequency, and a discharge warning is output. When the trajectory offset and the coefficient of variation of the current distribution width are both within the normal range, the power flow direction of the corresponding path segment in the multidimensional current propagation sequence is opposite to the theoretical power flow direction, and no change in the state of the switching device is detected downstream of the path segment, it is determined that there is a bypass leakage fault caused by insulation breakdown in the path segment, and the reverse tracing is performed to the nearest weak point in insulation.

[0048] When the transfer function distortion degree shows a significant anomaly, and the frequency domain characteristics of the corresponding path segment in the multidimensional current propagation sequence show a sudden increase in the amplitude of high-frequency components or a drastic change in harmonic content, it is determined that there is a partial discharge or arc fault in the path segment. The path segment where the discharge point is located is then located based on the peak value of the distortion frequency, and a discharge warning is output. Specifically, the transfer function distortion degree of the corresponding path segment is obtained. The distortion degree of the transfer function is compared with a preset distortion threshold. When comparing, When an abnormal change occurs in the current propagation relationship of the path segment, the preset distortion threshold is set to 0.3. Further, a Fast Fourier Transform is performed on the time-domain data of the current in the multidimensional current propagation sequence to obtain the spectral distribution result of the corresponding path segment. Extract the amplitude corresponding to each frequency component. and the proportion of harmonic energy When the amplitude of the high-frequency component within the frequency range of 10kHz to 100kHz satisfies: ,in, Indicates the amplitude of high-frequency components. This represents the average amplitude of the corresponding frequency band under normal operating conditions; and the proportion of harmonic energy satisfies: At that time, it was determined that there was a high-frequency abnormal current disturbance in the current path segment; further, the abnormal frequency component with the largest amplitude was determined based on the spectrum distribution results. The current propagation path segment corresponding to the abnormal frequency component is determined as the candidate region for discharge location, and spatial matching is performed by combining the path segment number corresponding to the maximum value of the transfer function distortion to determine the location of partial discharge or arc fault; the discharge point location and discharge warning information are output according to the spatial coordinates of the corresponding path segment.

[0049] When the trajectory offset and the coefficient of variation of the current distribution width are both within the normal range, the power flow direction of the corresponding path segment in the multidimensional current propagation sequence is opposite to the theoretical power flow direction, and no change in the state of the switching device is detected downstream of the path segment, it is determined that there is a bypass leakage fault caused by insulation breakdown in the path segment, and the fault is traced in reverse to the nearest weak point in the insulation; specifically, the trajectory offset of the corresponding path segment is... Offset threshold from preset trajectory Compare and use the coefficient of variation of the current distribution width Compared with the preset width variation threshold When comparing, and At that time, it was determined that the spatial distribution of current in the current path segment did not show any obvious abnormalities; further, based on the power flow direction characteristics in the multidimensional current propagation sequence... The theoretical power flow direction is determined based on the conductor connection structure and switching state. When comparing: When the actual current propagation direction is opposite to the theoretical power supply direction, the status of the downstream switching device in the corresponding path segment is read. When the switch status remains unchanged within the continuous detection cycle and there is no active power supply switching behavior, it is determined that the reverse current is not generated by switch control, but by bypass current formed by insulation abnormality.

[0050] Furthermore, taking the abnormal path segment where the power flow direction reverses as the tracing starting point, the transfer function distortion, current density change, and spatial magnetic field anomaly intensity of adjacent path segments are obtained sequentially in the opposite direction of the conduction path; the insulation anomaly evaluation value corresponding to each candidate path segment is calculated: ,in, This represents the insulation anomaly evaluation value of the i-th candidate path segment. This indicates the distortion degree of the transfer function. This represents the change in current density. Indicates the strength of the magnetic field anomaly. , , These represent the corresponding weights; in the specific implementation process, the following settings are made: , , .

[0051] According to the insulation anomaly evaluation value Sort the data from largest to smallest, identify the path segment with the largest evaluation value as the nearest weak point in insulation, and output the spatial coordinates of the weak point, the number of the conductive path to which it belongs, and the bypass leakage fault alarm information.

[0052] Furthermore, the process of performing distribution box fault matching based on consistency indicators and multidimensional current propagation sequences, and establishing online detection results, also includes: When the trajectory offset, current distribution width variation coefficient, and transfer function distortion are all within the normal range, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence shows that the current amplitude is significantly lower than the normal level and the fundamental amplitude in the frequency domain characteristics decreases synchronously, it is determined that there is an abnormal upstream power supply fault, and an alarm for phase loss or voltage drop is triggered.

[0053] Obtain the trajectory offset corresponding to each path segment Coefficient of variation of current distribution width and the distortion of the transfer function And compare them with the corresponding anomaly detection thresholds respectively; when the trajectory offset satisfies The coefficient of variation of the current distribution width satisfies And the distortion degree of the transfer function satisfies If the current spatial propagation position, current diffusion range, and current transmission relationship of the current path segment are determined to have not changed abnormally, conductor offset, contact abnormality, and local propagation distortion faults are ruled out.

[0054] Furthermore, the time-domain waveform of the corresponding path segment in the multidimensional current propagation sequence is periodically analyzed to obtain the current effective value. And read the reference current value of the corresponding path segment under normal power supply conditions. ; Calculate the current current amplitude attenuation ratio: ,in, This indicates the percentage decrease in current amplitude along the current path segment. This represents the reference current value for the corresponding path segment under normal power supply conditions. This represents the effective value of the current obtained in real time.

[0055] When the current amplitude attenuation ratio When the current amplitude is greater than a preset amplitude attenuation threshold, it is determined that the current path segment has a reduced power supply capacity; wherein, the preset amplitude attenuation threshold is set to 0.3, that is, when the real-time effective current value is lower than 70% of the normal operating current value, it is determined that the current amplitude has decreased significantly.

[0056] Furthermore, frequency domain transformation is performed on the multidimensional current propagation sequence to obtain the spectral response results of the corresponding path segments. Extract the amplitude corresponding to the fundamental frequency. and the corresponding fundamental amplitude under normal operating conditions Calculate the percentage change in the fundamental amplitude: ,in, This indicates the percentage decrease in the fundamental frequency amplitude. This represents the fundamental frequency amplitude under normal conditions. This indicates the fundamental amplitude under the current condition.

[0057] When the fundamental amplitude decreases by a certain percentage Greater than 0.3, and the attenuation ratio of the time-domain current amplitude When the synchronization exceeds 0.3, it is determined that there is an abnormality in the power supply side input of the current path segment; further, the relationship between the current amplitude changes of each phase of the corresponding path segment is judged. When the amplitude of a single phase current drops to less than 20% of the corresponding amplitude under normal conditions, while the current of other phases remains at a normal level, it is determined that there is a phase loss fault; when the amplitude of each phase current drops synchronously, and the proportion of the fundamental amplitude decreases but the frequency characteristics remain stable, it is determined that there is a voltage drop fault.

[0058] After determining that the upstream power supply is abnormal, the current amplitude changes of adjacent nodes are obtained by tracing back to the upstream node along the power supply direction according to the topology of the conduction path corresponding to the abnormal path segment. When the current amplitude of multiple consecutive downstream path segments decreases, while the current state of the path segment before the upstream abnormal node is normal, the upstream abnormal node is determined as the location affected by the power supply abnormality, and the corresponding path segment number, spatial coordinates, and phase loss or voltage drop alarm information are output.

[0059] In summary, the embodiments of this application have at least the following technical effects: First, based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, at least one current conduction path is determined, and each conduction path is divided into multiple continuous path segments according to physical connection nodes. Next, non-invasive current sensors synchronously collect current time-series data at the nodes of each path segment. Based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. These current propagation characteristics include at least time-domain waveforms, frequency-domain characteristics, and power flow direction, establishing a multi-dimensional current propagation sequence for each conduction path. Then, feedback signals from magnetic field sensor arrays deployed along the spatial distribution area of ​​the conduction path are read to establish a spatial magnetic field signal. The spatial partitioning of the inverse current density is reconstructed using magnetic field inverse sources, and the actual spatial trajectory of the current along each conduction path is extracted. Further, the multi-dimensional current propagation sequence is mapped to the spatial coordinates of the path segments and compared segment by segment with the actual spatial trajectory to establish a consistency index. Finally, based on the consistency index and the multi-dimensional current propagation sequence, distribution box fault matching is performed to establish online detection results. This invention solves the technical problem in existing technologies where the complex internal wiring structure of distribution boxes makes it difficult to identify abnormal current states in a timely and accurate manner, resulting in insufficient fault location accuracy. It achieves the technical effect of improving the accuracy and reliability of online fault detection and location in distribution boxes.

[0060] Example 2 is based on the same inventive concept as the IoT-based online fault detection method for distribution boxes in the previous examples, such as... Figure 4 As shown, this application provides an online fault detection system for distribution boxes based on the Internet of Things, wherein the system includes: Conductive path segmentation module 11: Based on the actual connection structure of conductors inside the distribution box and the on / off state of switching devices, determine at least one conductive path for the current, and divide each conductive path into multiple continuous path segments according to physical connection nodes; Current feature extraction module 12: Synchronously collect current time-series data at the nodes of each path segment using non-invasive current sensors, and extract the current propagation characteristics of each path segment based on the current transmission relationship between adjacent nodes on the same conductive path. The current propagation characteristics include at least time-domain waveform, frequency-domain characteristics, and power flow direction, establishing the current propagation characteristics of each conductive path. Multidimensional current propagation sequence; Current trajectory inversion module 13: Reads the feedback signal of the magnetic field sensor array deployed in the spatial distribution area along the conduction path, establishes the spatial magnetic field signal, reconstructs the spatial partition of the inverted current density using the magnetic field inverse source, and extracts the actual spatial trajectory of the current along each conduction path; Comparison module 14: Maps the multidimensional current propagation sequence to the spatial coordinates of the path segment, and compares it segment by segment with the actual spatial trajectory to establish a consistency index; Fault matching module 15: Performs distribution box fault matching based on the consistency index and the multidimensional current propagation sequence to establish online detection results.

[0061] Furthermore, the comparison module 14 is used to perform the following method: The consistency index includes at least one of trajectory offset, current distribution width variation coefficient, and transfer function distortion.

[0062] Furthermore, the comparison module 14 is used to perform the following method: The amplitude envelope and zero-crossing sequence of the current time-domain waveform of each path segment are extracted from the multidimensional current propagation sequence to establish the behavioral fingerprint of the current in the time domain for the corresponding path segment. Based on the amplitude envelope change rate and zero-crossing density distribution in the behavioral fingerprint, an equivalent current distribution function characterizing the intensity of the current distribution in the cross-section of the path segment is constructed, and the equivalent current distribution function is continuously expanded along the extension direction of the path segment to form a current spatial distribution expression. The current spatial distribution expression is weighted and integrated in the direction perpendicular to the conduction path, and the connected trajectory of the region with the maximum current density is extracted as the current centerline. Based on the spatial coordinates, the current centerline is aligned to the corresponding spatial position of the actual spatial trajectory according to the timestamp, and the offset distance sequence between the current centerline and the spatial trajectory centerline in the direction perpendicular to the path is calculated. The root mean square value of the offset distance sequence is output as the trajectory offset. When a non-periodic abrupt change occurs in the time domain waveform and the actual spatial trajectory does not change accordingly, it is determined that the current behavior is mismatched with the actual spatial trajectory distribution, and an unexpected conduction result is output.

[0063] Furthermore, the comparison module 14 is used to perform the following method: Based on the amplitude envelope in the multidimensional current propagation sequence, the lateral current diffusion range corresponding to each path segment is determined, and a current distribution width reference region is established according to the spatial boundary position of the corresponding path segment in the actual spatial trajectory. The lateral current diffusion range of each path segment is projected onto the spatial coordinate system where the actual spatial trajectory is located, and the overlap width and deviation width between the current distribution range and the current distribution width reference region are calculated based on the projection results to form a current distribution width sequence under spatial constraints. The current distribution width sequences of adjacent path segments are continuously compared along the conduction path direction, and the dispersion of the current distribution width of each path segment relative to the width of the actual spatial trajectory is calculated to construct the current distribution width variation coefficient.

[0064] Furthermore, the comparison module 14 is used to perform the following method: Based on the current timing data of adjacent path segments on the same conduction path, the amplitude transfer relationship and phase transfer relationship between adjacent nodes are obtained, and the current transfer characteristics corresponding to the path segment are established. According to the spatial continuity relationship between adjacent path segments in the actual spatial trajectory, the spatial transfer range corresponding to the current in the normal conduction state is determined, and a reference transfer characteristic matching the spatial transfer range is established. The current transfer characteristics and the reference transfer characteristics are compared on a path segment by path, and the amplitude attenuation deviation, phase delay deviation and spatial position offset during the actual current propagation process are extracted and fused to form the transfer function distortion degree.

[0065] Furthermore, the fault matching module 15 is used to perform the following method: When the trajectory offset and the coefficient of variation of the current distribution width both show significant anomalies, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence exhibits a non-periodic abrupt change and amplitude envelope attenuation, it is determined that there is an abnormal contact resistance fault in the path segment, and the spatial coordinates and severity level of the path segment are output.

[0066] Furthermore, the fault matching module 15 is used to perform the following method: When the distortion of the transfer function is significantly abnormal, and the frequency domain characteristics of the corresponding path segment in the multidimensional current propagation sequence show a sudden increase in the amplitude of high-frequency components or a drastic change in harmonic content, it is determined that there is a partial discharge or arc fault in the path segment, and the path segment where the discharge point is located is located based on the peak value of the distortion frequency, and a discharge warning is output; when the trajectory offset and the coefficient of variation of the current distribution width are both within the normal range, the power flow direction of the corresponding path segment in the multidimensional current propagation sequence is opposite to the theoretical power flow direction, and no change in the state of the switching device is detected downstream of the path segment, it is determined that there is a bypass leakage fault caused by insulation breakdown in the path segment, and the fault is traced in reverse to the nearest weak insulation point.

[0067] Furthermore, the fault matching module 15 is used to perform the following method: When the trajectory offset, current distribution width variation coefficient, and transfer function distortion are all within the normal range, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence shows that the current amplitude is significantly lower than the normal level and the fundamental amplitude in the frequency domain characteristics decreases synchronously, it is determined that there is an abnormal upstream power supply fault, and an alarm for phase loss or voltage drop is triggered.

[0068] Example 3: Based on the same inventive concept as the IoT-based online fault detection method for distribution boxes in the preceding examples, this example provides a computer-readable storage medium for storing software programs, computer-executable programs, and modules, such as the program instructions / modules corresponding to the IoT-based online fault detection method for distribution boxes in this application. The processor executes the software programs, instructions, and modules stored in the memory to perform various functional applications and data processing of the computer device, thereby realizing the aforementioned IoT-based online fault detection method for distribution boxes.

[0069] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A method for online fault detection in distribution boxes based on the Internet of Things, characterized in that, The method includes: Based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, determine at least one conduction path for the current, and divide each conduction path into multiple continuous path segments according to the physical connection nodes. At the nodes of each path segment, current time-series data are synchronously collected using non-invasive current sensors. Based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation characteristics of each path segment are extracted. The current propagation characteristics include at least time-domain waveform, frequency-domain characteristics, and power flow direction, and a multi-dimensional current propagation sequence for each conduction path is established. Read the feedback signal of the magnetic field sensor array deployed in the spatial distribution area along the conduction path, establish the spatial magnetic field signal, reconstruct the spatial partition of the current density using the magnetic field inverse source, and extract the actual spatial trajectory of the current along each conduction path; The multidimensional current propagation sequence is mapped to the spatial coordinates of the path segment and compared with the actual spatial trajectory segment by segment to establish a consistency index. Based on the consistency index and multidimensional current propagation sequence, fault matching of distribution boxes is performed, and online detection results are established.

2. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 1, characterized in that, The consistency index includes at least one of trajectory offset, current distribution width variation coefficient, and transfer function distortion.

3. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, The trajectory offset is calculated based on a point-by-point comparison between the time-domain waveform and the actual spatial trajectory, including: The amplitude envelope and zero-crossing sequence of the current time domain waveform of each path segment are extracted from the multidimensional current propagation sequence to establish the behavioral fingerprint of the current in the time domain of the corresponding path segment. Based on the amplitude envelope change rate and zero-crossing density distribution in the behavioral fingerprint, an equivalent current distribution function characterizing the intensity of current distribution in the cross section of the path segment is constructed, and the equivalent current distribution function is continuously expanded along the extension direction of the path segment to form a spatial distribution expression of current. The current spatial distribution is weighted and integrated in the direction perpendicular to the conduction path, and the connected trajectory of the region with the maximum current density is extracted as the current centerline. Based on the spatial coordinates, the current centerline is aligned to the corresponding spatial position of the actual spatial trajectory according to the timestamp. The offset distance sequence between the current centerline and the spatial trajectory centerline in the vertical path direction is calculated. The root mean square value of the offset distance sequence is output as the trajectory offset. When a non-periodic abrupt change occurs in the time domain waveform and the actual spatial trajectory does not change accordingly, it is determined that the current behavior is mismatched with the distribution of the actual spatial trajectory, and an unexpected conduction result is output.

4. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, The calculation of the coefficient of variation of the current distribution width includes: The lateral diffusion range of the current corresponding to each path segment is determined based on the amplitude envelope in the multidimensional current propagation sequence, and a reference area for the current distribution width is established according to the spatial boundary position of the corresponding path segment in the actual spatial trajectory. Project the lateral current diffusion range of each path segment onto the spatial coordinate system where the actual spatial trajectory is located, and calculate the overlap width and deviation width between the current distribution range and the current distribution width reference area based on the projection results to form a current distribution width sequence under spatial constraints. The current distribution width sequence of adjacent path segments is continuously compared along the conduction path direction. The dispersion of the current distribution width of each path segment relative to the actual spatial trajectory width is calculated, and the current distribution width variation coefficient is constructed.

5. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, The calculation of the transfer function distortion includes: Based on the current timing data of adjacent path segments on the same conduction path, the amplitude transfer relationship and phase transfer relationship between adjacent nodes are obtained, and the current transfer characteristics corresponding to the path segment are established. Based on the spatial continuity relationship between adjacent path segments in the actual spatial trajectory, the spatial transmission range corresponding to the current in the normal conduction state is determined, and a reference transmission feature matching the spatial transmission range is established. The current transmission characteristics are compared with the reference transmission characteristics on a path segment by path segment to extract the amplitude attenuation deviation, phase delay deviation and spatial position offset during the actual current propagation process, and then fused to form the transfer function distortion degree.

6. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, Based on consistency indicators and multidimensional current propagation sequences, fault matching of distribution boxes is performed, and online detection results are established, including: When the trajectory offset and the coefficient of variation of the current distribution width both show significant anomalies, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence exhibits a non-periodic abrupt change and amplitude envelope attenuation, it is determined that there is an abnormal contact resistance fault in the path segment, and the spatial coordinates and severity level of the path segment are output.

7. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, Based on consistency indicators and multidimensional current propagation sequences, distribution box fault matching is performed, and online detection results are established. This also includes: When the distortion of the transfer function is significantly abnormal, and the frequency domain characteristics of the corresponding path segment in the multidimensional current propagation sequence show a sudden increase in the amplitude of high-frequency components or a drastic change in harmonic content, it is determined that there is a partial discharge or arc fault in the path segment, and the path segment where the discharge point is located is located based on the peak value of the distortion frequency, and a discharge warning is output. When the trajectory offset and the coefficient of variation of the current distribution width are both within the normal range, the power flow direction of the corresponding path segment in the multidimensional current propagation sequence is opposite to the theoretical power flow direction, and no change in the state of the switching device is detected downstream of the path segment, it is determined that there is a bypass leakage fault caused by insulation breakdown in the path segment, and the reverse tracing is performed to the nearest weak point in insulation.

8. The method for online fault detection of distribution boxes based on the Internet of Things as described in claim 2, characterized in that, Based on consistency indicators and multidimensional current propagation sequences, distribution box fault matching is performed, and online detection results are established. This also includes: When the trajectory offset, current distribution width variation coefficient, and transfer function distortion are all within the normal range, and the time domain waveform of the corresponding path segment in the multidimensional current propagation sequence shows that the current amplitude is significantly lower than the normal level and the fundamental amplitude in the frequency domain characteristics decreases synchronously, it is determined that there is an abnormal upstream power supply fault, and an alarm for phase loss or voltage drop is triggered.

9. An online fault detection system for distribution boxes based on the Internet of Things, characterized in that, For implementing the Internet of Things-based online fault detection method for distribution boxes according to any one of claims 1-8, the system comprises: Conductive path division module: Based on the actual connection structure of the conductors inside the distribution box and the on / off state of the switching devices, determine at least one conductive path for the current, and divide each conductive path into multiple continuous path segments according to the physical connection nodes; Current feature extraction module: At the nodes of each path segment, current time series data is synchronously collected by non-invasive current sensors, and based on the current transmission relationship between adjacent nodes on the same conduction path, the current propagation features of each path segment are extracted. The current propagation features include at least time domain waveform, frequency domain characteristics and power flow direction, and a multidimensional current propagation sequence for each conduction path is established. Current trajectory inversion module: Reads the feedback signal of the magnetic field sensor array deployed in the spatial distribution area along the conduction path, establishes the spatial magnetic field signal, reconstructs the spatial partition of the current density using the magnetic field inverse source, and extracts the actual spatial trajectory of the current along each conduction path; Comparison module: Maps the multidimensional current propagation sequence to the spatial coordinates of the path segment, and compares it segment by segment with the actual spatial trajectory to establish a consistency index; Fault matching module: Performs fault matching of distribution boxes based on consistency indicators and multidimensional current propagation sequences, and establishes online detection results.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the Internet of Things-based online fault detection method for distribution boxes as described in any one of claims 1-8.

Images