Distribution cable fault processing method and device, electronic equipment and storage medium

Abstract

The invention discloses a power distribution cable fault processing method and device, electronic equipment and a storage medium, and relates to the technical field of power distribution cable fault processing, and the method comprises the steps: firstly collecting multi-mode fault signals, such as a high-frequency traveling wave signal, an ultrasonic partial discharge signal and an environment parameter signal, of a power distribution cable; and distribution network topological structure information such as line length, branch number, node position, wave impedance parameters and the like is synchronously obtained. Then high-frequency traveling wave features are extracted, a positioning strategy is determined according to the number of line branches, and fault position coordinates are obtained in cooperation with the traveling wave features and topological information. Then partial discharge features and environment features are extracted respectively, the three types of features are fused to form a fusion feature set, the feature set is analyzed through a fault classification model optimized by the improved whale algorithm, and fault type information is obtained. And finally, associating fault positions and types, generating a fault processing scheme containing a positioning result, fault causes and operation and maintenance priorities, and realizing accurate positioning, classification and operation and maintenance guidance of the distribution cable faults.

Description

Technical Field

[0001] This application relates to the field of power distribution cable fault handling technology, and in particular to a power distribution cable fault handling method, device, electronic device and storage medium. Background Technology

[0002] Power distribution cable networks (especially 10kV distribution networks) are characterized by numerous branches and complex topologies. When a fault occurs, traveling wave signals are reflected and refracted between multiple nodes, leading to feature confusion and posing a significant challenge to fault location and type identification. Traveling wave theory serves as the core basis for fault detection, and its propagation characteristics (such as wave velocity, attenuation, and reflection coefficient) contain rich fault information. However, existing technologies have limitations in adapting to complex topologies, distinguishing between multiple fault types, and fusing data from multiple terminals.

[0003] Currently, fault location relies heavily on graph theory or single-ended traveling wave methods, while fault identification primarily uses traditional machine learning. However, the two lack a collaborative mechanism. In complex distribution networks, graph theory-based location methods experience a significant increase in algorithm complexity when there are many branches, while traditional machine learning models have insufficient accuracy in identifying similar faults such as partial discharge and lightning strikes, making it difficult to meet the needs of efficient operation and maintenance of distribution networks. Summary of the Invention

[0004] In view of this, this application provides a method, apparatus, electronic device and storage medium for handling power distribution cable faults, which can improve the accuracy of fault identification of power distribution cables and meet the needs of efficient operation and maintenance of power distribution networks.

[0005] According to a first aspect of this application, a method for handling power distribution cable faults is provided, comprising:

[0006] The system collects multi-mode fault signals from power distribution cables and simultaneously acquires distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location, and wave impedance parameters of each line segment.

[0007] Feature extraction is performed on the high-frequency traveling wave signal to obtain traveling wave features. A fault location strategy is determined based on the number of line branches in the distribution network topology information. The traveling wave features and the distribution network topology information are used to locate the fault and obtain the fault location coordinates of the distribution cable.

[0008] Feature extraction is performed on the ultrasonic partial discharge signal and the environmental parameter signal to obtain partial discharge features and environmental features. The traveling wave features, the partial discharge features and the environmental features are fused to form a fused feature set. The fused feature set is analyzed based on the fault classification model optimized by the improved whale algorithm to classify similar faults and obtain the fault type information of the power distribution cable.

[0009] By associating the fault location coordinates with the fault type information, a fault handling plan is generated that includes the location result, the cause of the fault, and the maintenance priority.

[0010] According to a second aspect of this application, a power distribution cable fault handling device is provided, comprising:

[0011] The acquisition module is used to acquire multi-mode fault signals of power distribution cables and simultaneously obtain distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location and wave impedance parameters of each line segment.

[0012] The positioning module is used to extract features from the high-frequency traveling wave signal to obtain traveling wave features, determine the fault location strategy based on the number of line branches in the distribution network topology information, and coordinate the traveling wave features with the distribution network topology information to perform fault location and obtain the fault location coordinates of the distribution cable.

[0013] The classification module is used to extract features from the ultrasonic partial discharge signal and the environmental parameter signal respectively to obtain partial discharge features and environmental features. The traveling wave features, partial discharge features and environmental features are fused to form a fused feature set. The fused feature set is analyzed based on the fault classification model optimized by the improved whale algorithm to classify similar faults and obtain the fault type information of the power distribution cable.

[0014] The generation module is used to associate the fault location coordinates with the fault type information to generate a fault handling plan that includes the location result, fault cause and maintenance priority.

[0015] According to a third aspect of this application, a storage medium is provided that stores a computer program thereon, which, when executed by a processor, implements the above-described power distribution cable fault handling method.

[0016] According to a fourth aspect of this application, an electronic device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described power distribution cable fault handling method.

[0017] By employing the aforementioned technical solutions, the power distribution cable fault handling method, apparatus, electronic equipment, and storage medium provided in this application, through the acquisition of multi-mode fault signals composed of high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals, and simultaneously acquiring distribution network topology information such as line length and number of branches, can overcome the limitations of traditional fault handling methods that rely on a single signal dimension and insufficient utilization of topology information, providing multi-dimensional and precise data support for subsequent location and classification. Secondly, after extracting high-frequency traveling wave features, the fault location strategy is dynamically determined based on the number of line branches in the distribution network topology, and location is performed in conjunction with the traveling wave features and topology information. This approach not only adapts differentiated algorithms to different branch number scenarios to reduce complexity, but also improves fault location accuracy through the combination of features and topology. The improved positioning accuracy and efficiency of coordinate positioning can solve the problem of the drastically increased complexity of traditional graph theory positioning algorithms in complex distribution networks. Furthermore, after extracting partial discharge features and environmental features, the system integrates traveling wave features to form a multi-dimensional fusion feature set. This set is then combined with an improved fault classification model optimized by the whale algorithm for analysis. Through multi-feature complementarity and model parameter optimization, the classification accuracy of similar faults such as partial discharge and lightning strikes can be effectively improved, making up for the shortcomings of traditional machine learning in identification. Finally, by associating fault location coordinates with fault type information, a processing solution containing positioning results, fault causes, and maintenance priorities is generated. This breaks down the barriers between traditional fault location and identification, forming a complete collaborative mechanism for data collection, positioning, classification, and maintenance guidance, which can fully meet the needs of efficient operation and maintenance in complex distribution networks.

[0018] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0019] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings:

[0020] Figure 1 A flowchart illustrating a power distribution cable fault handling method provided in an embodiment of this application is shown.

[0021] Figure 2 A flowchart illustrating a power distribution cable fault handling method according to another embodiment of this application is shown;

[0022] Figure 3 A schematic diagram of the structure of a power distribution cable fault handling device provided in an embodiment of this application is shown. Detailed Implementation

[0023] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0024] Currently, fault location relies heavily on graph theory or single-ended traveling wave methods, while fault identification primarily uses traditional machine learning. However, the two lack a collaborative mechanism. In complex distribution networks, graph theory-based location methods experience a significant increase in algorithm complexity when there are many branches, while traditional machine learning models have insufficient accuracy in identifying similar faults such as partial discharge and lightning strikes, making it difficult to meet the needs of efficient operation and maintenance of distribution networks.

[0025] To address the aforementioned technical problems, embodiments of the present invention provide a method for handling power distribution cable faults, such as... Figure 1 As shown, the method includes:

[0026] Step 110: Collect multi-mode fault signals of power distribution cables and simultaneously acquire distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location, and wave impedance parameters of each line segment.

[0027] Multimodal fault signals refer to a collection of various signals acquired from different dimensions to characterize the fault state of distribution cables. These signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals, comprehensively reflecting the transient characteristics, local defects, and external influencing factors of the fault. High-frequency traveling wave signals refer to the high-frequency transient traveling waves generated in the distribution cable when a fault occurs. Their core function is to capture the propagation characteristics of transient faults such as lightning strikes, phase-to-phase short circuits, and open circuits, providing transient data support for fault location. Ultrasonic partial discharge signals refer to the ultrasonic signals (typically tens to hundreds of kHz) generated when partial discharge occurs inside the cable insulation. Their core function is to monitor latent defects such as insulation aging and air gap breakdown, providing local defect characteristics for fault classification. Environmental parameter signals refer to the physical parameter signals of the distribution cable's operating environment, mainly including temperature and humidity. The core function of parameters such as humidity level is to reflect the impact of the external environment on cable faults (e.g., high humidity accelerates insulation aging), providing an environmental basis for fault cause analysis; the distribution network topology information refers to a set of parameters reflecting the physical structure and electrical characteristics of the distribution network, and its core function is to provide a structural and electrical foundation for fault location; the line length is the actual physical length of each section of the distribution cable, used to calculate the traveling wave propagation time to locate the fault; the number of line branches is the total number of line branches in the distribution network, used to determine the complexity of the distribution network and select an appropriate fault location strategy; the node location is the geographical coordinate or relative position of the line connection point (such as towers, switchgear) in the distribution network, used to anchor the fault spatial range; the wave impedance parameters of each section of the line are the characteristic impedances of each section of the distribution cable (determined by the line material, cross-section, etc.), used to calculate the traveling wave reflection coefficient to assist in determining the fault location.

[0028] By acquiring multimodal fault signals, the limitations of traditional single-signal methods (such as high-frequency traveling waves) in comprehensively characterizing faults can be overcome. High-frequency traveling wave signals can capture transient faults such as lightning strikes and short circuits; ultrasonic partial discharge signals can identify latent defects such as insulation aging; and environmental parameter signals can correlate the impact of external factors such as temperature and humidity on faults. The combination of these three methods enables a comprehensive description of the fault. Simultaneously acquired distribution network topology information provides crucial physical and electrical basis for subsequent fault location, avoiding the problems of poor strategy adaptability and large location errors caused by the lack of topology parameters (such as the number of branches and wave impedance) in traditional fault location methods.

[0029] Step 120: Extract features from the high-frequency traveling wave signal to obtain traveling wave features. Determine the fault location strategy based on the number of line branches in the distribution network topology information. Combine the traveling wave features with the distribution network topology information to locate the fault and obtain the fault location coordinates of the distribution cable.

[0030] Among them, traveling wave features are quantitative parameters extracted from high-frequency traveling wave signals that characterize the transient characteristics of faults. They typically include wavelet energy spectrum (reflecting the energy distribution of traveling waves in different frequency bands) and traveling wave reflection coefficient (reflecting the reflection pattern of traveling waves at different wave impedance line connections). The fault location strategy is a fault location calculation method selected based on the complexity of the distribution network. For fewer branches, efficient methods such as simplified graph theory are used, while for more branches, high-precision methods such as multi-terminal traveling wave fusion are adopted, aiming to balance location efficiency and accuracy. Fault location coordinates refer to the specific spatial coordinates (geographic coordinates or coordinates relative to distribution network nodes) of the fault point of the distribution cable, calculated by coordinating traveling wave features and distribution network topology information, providing accurate location guidance for subsequent operation and maintenance.

[0031] In this embodiment of the present disclosure, the features of the acquired high-frequency traveling wave signal can be extracted first. Typically, the wavelet energy spectrum is calculated by wavelet transform (e.g., using the db6 wavelet basis for 8-level decomposition), and the traveling wave reflection coefficient is derived by combining the wave impedance parameters in the distribution network topology. Together, they constitute the traveling wave features that can characterize the transient features of the fault. Next, the fault location strategy is selected based on the number of line branches in the distribution network topology information: if the number of branches is small (≤ a preset threshold), a low-complexity method such as simplified graph theory is used; if the number of branches is large (> a preset threshold), a method adapted to complex distribution networks, such as multi-terminal traveling wave fusion, is used. Finally, the extracted traveling wave features are calculated in conjunction with the distribution network topology information. By analyzing the matching relationship between the arrival time of the traveling wave front, the reflection law, and the line structure, the fault location coordinates of the distribution cable are finally determined.

[0032] Extracting traveling wave features provides microscopic signal evidence for fault location, avoiding feature ambiguity caused by relying solely on the original signal. Dynamically selecting a location strategy based on the number of line branches improves location efficiency in simple distribution networks by simplifying methods, and reduces algorithm complexity in complex distribution networks by using multi-terminal fusion and other methods, thus solving the problem of performance degradation of traditional fixed strategies when the number of branches increases. The collaborative calculation of traveling wave features and distribution network topology deeply integrates the signal characteristics of the fault with the physical structure of the line, which can significantly improve the accuracy and anti-interference capability of location, effectively reduce location errors caused by electromagnetic interference and complex line structures, and provide accurate location support for subsequent fault handling.

[0033] Step 130: Extract features from the ultrasonic partial discharge signal and environmental parameter signal respectively to obtain partial discharge features and environmental features. Fuse the traveling wave features, partial discharge features and environmental features to form a fused feature set. Analyze the fused feature set based on the improved whale algorithm-optimized fault classification model to classify similar faults and obtain fault type information of the power distribution cable.

[0034] Among them, partial discharge features are quantitative indicators extracted from ultrasonic partial discharge signals, typically including the maximum amplitude, average amplitude, and repetition frequency of ultrasonic pulses per unit time. Their core function is to characterize the local defect state of cable insulation and assist in distinguishing faults caused by insulation problems. Environmental features are statistical features extracted from environmental parameter signals, often including extreme values ​​and trend vectors of parameters such as temperature and humidity. Their core function is to reflect the impact of environmental factors on faults (e.g., high humidity accelerates insulation aging), providing auxiliary basis for classifying similar faults. The fusion feature set is a comprehensive feature set formed by splicing traveling wave features, partial discharge features, and environmental features according to dimensions. Its core characteristic is multi-dimensional interoperability. The supplementary algorithm comprehensively characterizes the transient, defective, and environmental attributes of faults, providing sufficient feature support for the classification of similar faults. The improved whale algorithm is an improved version of the traditional whale optimization algorithm. Its core is to enhance the parameter optimization capability by adjusting the convergence strategy, which is used to optimize the key parameters of the fault classification model and improve the model classification performance. The fault classification model is an algorithm model built based on machine learning or deep learning (such as LightGBM, neural networks). Its core function is to learn and fuse the mapping relationship between feature sets and fault types to achieve automatic identification of fault types. Fault type information refers to the specific category of power distribution cable faults determined by the model classification, such as partial discharge, lightning strike, insulation aging, phase-to-phase short circuit, etc.

[0035] In this embodiment of the disclosure, for ultrasonic partial discharge signals, after filtering, quantitative indicators such as the maximum amplitude, average amplitude, and repetition frequency of the pulse per unit time are statistically analyzed to form partial discharge characteristics. For environmental parameter signals, the maximum and minimum values ​​and trend vectors of parameters such as temperature and humidity are calculated through sliding window processing to form environmental characteristics. Then, the previously extracted traveling wave characteristics are concatenated with the newly obtained partial discharge characteristics and environmental characteristics according to dimensions to integrate them into a fusion feature set covering fault transients, local defects, and environmental impacts. Finally, a fault classification model (such as LightGBM) with optimized parameters based on the improved whale algorithm is used to analyze the fusion feature set. By learning the mapping relationship between features and fault types through the model, the model focuses on achieving accurate classification of similar faults such as partial discharge and lightning strikes, and outputs the final fault type information.

[0036] This embodiment addresses the core pain points of traditional fault identification methods—namely, the one-sidedness of features and the difficulty in distinguishing similar faults—through multi-feature fusion and optimized model classification design. Firstly, the fusion of traveling wave features, partial discharge features, and environmental features enables multi-dimensional information complementarity. Traveling wave features focus on the propagation characteristics of transient faults, partial discharge features capture latent insulation defects, and environmental features relate to external influencing factors. The combination of these three features avoids the limitations of a single feature in fault characterization, significantly improving the ability to identify similar faults. Secondly, the improved whale algorithm, by optimizing key parameters of the fault classification model (such as learning step size and model complexity), balances the model's global search and local optimization capabilities, avoiding overfitting or underfitting problems caused by inappropriate parameters in traditional machine learning models, and further enhancing the model's ability to capture subtle feature differences. The synergistic effect of these two approaches significantly improves the classification accuracy of similar faults such as partial discharge and lightning strikes, providing accurate type criteria for subsequent fault handling solutions and effectively supporting efficient operation and maintenance of distribution networks.

[0037] Step 140: Associate the fault location coordinates with the fault type information to generate a fault handling plan that includes the location results, fault cause, and maintenance priority.

[0038] In this embodiment of the disclosure, the previously determined fault location coordinates can first be associated with the fault type information to clarify "where a fault occurred"; then, based on the association results, combined with the distribution network topology information, environmental characteristics, and typical causal patterns of fault types, the root cause of the fault can be deduced; at the same time, the maintenance priority is set according to the urgency of the fault type (e.g., phase-to-phase short circuit has a higher priority than minor partial discharge), the scope of impact (e.g., main line fault has a higher priority than branch line fault) and the difficulty of handling; finally, the location results (i.e., fault location), fault cause, and maintenance priority are integrated into a standardized fault handling plan to provide clear action guidance for distribution network maintenance personnel.

[0039] By combining "information association + full-element solution generation," the limitations of traditional fault handling—such as the separation of location and identification and the lack of practical guidance—can be completely overcome, achieving seamless integration from technical testing to operation and maintenance execution. The association between fault location and type avoids blind operation and maintenance based on knowing only the location but not the nature of the fault, or only knowing the fault type but not the location, laying the foundation for accurate handling. Deducing the cause of the fault based on multi-dimensional information helps operation and maintenance personnel solve problems at their root (e.g., optimizing line parameters for "impedance mismatch," rather than just dealing with surface faults), reducing the fault recurrence rate. Setting operation and maintenance priorities quantifies the urgency and impact of faults, rationally allocating manpower, equipment, and other operation and maintenance resources, avoiding resource waste (e.g., prioritizing lightning strike faults that pose a risk of power outages, and delaying the handling of minor partial discharges). The final generated complete solution transforms abstract technical data into executable operation and maintenance instructions, significantly improving operation and maintenance efficiency and targeting, effectively solving the pain points of slow fault handling, difficulty in locating root causes, and resource mismatch in complex distribution networks.

[0040] In summary, the power distribution cable fault handling method provided by this invention, by collecting multi-mode fault signals composed of high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals, and simultaneously acquiring distribution network topology information such as line length and number of branches, can overcome the limitations of traditional fault handling methods that rely on a single signal dimension and insufficient utilization of topology information, providing multi-dimensional and precise data support for subsequent location and classification. Secondly, after extracting high-frequency traveling wave features, the fault location strategy is dynamically determined based on the number of line branches in the distribution network topology, and location is performed in conjunction with traveling wave features and topology information. This approach not only adapts differentiated algorithms to different branch number scenarios to reduce complexity, but also improves the location accuracy of fault coordinates through the combination of features and topology. In terms of efficiency, it can solve the problem of the dramatic increase in complexity of traditional graph theory-based localization algorithms in complex distribution networks. Furthermore, after extracting partial discharge features and environmental features, it fuses traveling wave features to form a multi-dimensional fusion feature set. Combined with the improved whale algorithm-optimized fault classification model for analysis, through multi-feature complementarity and model parameter optimization, it can effectively improve the classification accuracy of similar faults such as partial discharge and lightning strikes, making up for the shortcomings of traditional machine learning in identification. Finally, by associating fault location coordinates with fault type information, it generates a processing plan that includes location results, fault causes, and maintenance priorities. This can break down the barriers between traditional fault location and identification, forming a complete collaborative mechanism for data collection, location, classification, and maintenance guidance, which can fully meet the needs of efficient operation and maintenance of complex distribution networks.

[0041] Furthermore, as a refinement and extension of the specific implementation of the above embodiments, and to fully illustrate the implementation of this embodiment, this embodiment also provides another method for handling power distribution cable faults, such as... Figure 2 As shown, the method includes:

[0042] Step 210: Collect multi-mode fault signals of power distribution cables and simultaneously acquire distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location, and wave impedance parameters of each line segment.

[0043] Step 220: Extract features from the high-frequency traveling wave signal to obtain traveling wave features. Determine the fault location strategy based on the number of line branches in the distribution network topology information. Combine the traveling wave features with the distribution network topology information to locate the fault and obtain the fault location coordinates of the distribution cable.

[0044] In the embodiments of this disclosure, when extracting features from a high-frequency traveling wave signal to obtain traveling wave features, step 220 may specifically include: selecting a db6 wavelet basis as the decomposition basis function, performing an 8-level discrete wavelet transform on the high-frequency traveling wave signal to obtain the high-frequency detail coefficients of the 1st to 8th levels, calculating the wavelet energy spectrum of each level based on the high-frequency detail coefficients; calculating the traveling wave reflection coefficient based on the wave impedance parameters of each line segment in the distribution network topology information; and determining the wavelet energy spectrum and the traveling wave reflection coefficient as the traveling wave features.

[0045] Choosing the db6 wavelet basis avoids the problem of insufficient capture of traveling wave transient signals by traditional wavelet basis (such as db4). Its good time-frequency resolution can accurately separate traveling wave components of different frequency bands. The 8-level discrete wavelet transform can cover the main frequency bands of the traveling wave signal, ensuring that key transient features are not missed by the high-frequency detail coefficients. The wavelet energy spectrum calculated based on the high-frequency detail coefficients can quantify the frequency domain energy differences of different faults (such as lightning strikes and short circuits), providing a basis for preliminary judgment of fault type. The traveling wave reflection coefficient, combined with the distribution network wave impedance parameters, is directly related to the reflection pattern of the traveling wave at line branches and nodes, which can provide a basis for locating the fault location based on propagation characteristics. The integrated traveling wave features take into account both frequency domain energy and propagation reflection dimensions, avoiding the shortcomings of insufficient information from traditional single features (such as using only wavefront time), and significantly improving the accuracy and anti-interference capability of subsequent fault location.

[0046] Accordingly, when determining the fault location strategy based on the number of line branches in the distribution network topology information, and performing fault location based on traveling wave characteristics and distribution network topology information to obtain the fault location coordinates of the distribution cable, step 220 of the embodiment may specifically include: if the number of line branches is less than or equal to a preset threshold, a simplified graph theory location method is adopted, constructing a distribution network topology node graph with node positions as vertices and line lengths as edges, marking the first arrival time of the traveling wave front collected by the high-frequency traveling wave sensor at each node, and calculating the fault location coordinates using Dijkstra's shortest path algorithm; if the number of line branches is greater than the preset threshold, a multi-terminal traveling wave information fusion location method is adopted, based on the distribution network topology... The line segments in the structural information are segmented, and the arrival times of the traveling wave fronts (TWF) collected by high-frequency traveling wave sensors at the beginning and end nodes of each segment are extracted. Using the line segment number as the row vector and the TWF arrival time as the column vector, time feature matrices are established at the beginning and end of the line segment, respectively. The time feature matrices at the beginning and end of the line segment are then fitted with the fault time matrix using least squares fitting. By minimizing the sum of squared errors between the actual arrival time and the theoretical arrival time, the TWF arrival time deviation caused by electromagnetic interference is eliminated, and the line segment where the fault is located is determined. The line segment is adjusted based on the reflection coefficient in the traveling wave features, and the fault location coordinates of the distribution cable are determined based on the adjusted line segment.

[0047] Among them, the preset threshold is a critical value for the number of branches (e.g., 3) used to determine the complexity of the distribution network, triggering different positioning strategies to balance positioning efficiency and accuracy; the simplified graph theory positioning method is a positioning method suitable for simple distribution networks, which abstracts the distribution network into a "node-edge" model based on graph theory and uses path algorithms to calculate the fault location, characterized by simple calculation and high efficiency; the distribution network topology node graph is a graphical model constructed with distribution network node locations as vertices and line lengths as edges, intuitively reflecting the physical connection relationship of the distribution network, and is the basic carrier of simplified graph theory positioning; the first arrival time of the traveling wave wavehead is the first detection time of the traveling wave from the fault point to each node sensor, containing key spatiotemporal information of the fault location, and is the core time series data for positioning calculation; Dijkstra's shortest path algorithm is an algorithm used to find the shortest path between two points in the graph, here calculating the shortest propagation path of the traveling wave from the fault point to each node (corresponding to the minimum time difference) to infer the fault location coordinates; multi-terminal traveling wave information fusion positioning... The positional method is a positioning method suitable for complex distribution networks. It integrates the traveling wave time information of multiple nodes to eliminate interference and improve accuracy. It is characterized by strong anti-interference ability and adaptability to multi-branch scenarios. The first / last end time feature matrix is ​​a matrix constructed with line segments as rows and the arrival time of the traveling wave fronts of the corresponding first / last end nodes as columns. It centrally carries the time sequence characteristics of traveling waves at multiple ends, which is convenient for batch analysis and fitting. Least square fitting is a mathematical method that achieves the optimal matching between data and model by minimizing the sum of squared errors between actual data and theoretical models. Here, it is used to eliminate the traveling wave arrival time deviation caused by electromagnetic interference and improve the reliability of time data. The fault time matrix is ​​a theoretical traveling wave arrival time matrix calculated based on line length and traveling wave propagation speed. It is used as a benchmark to fit with the actual time matrix and is used to calibrate errors. The line segment range where the fault is located is the line segment range where the fault may exist initially determined by fitting the time matrix. It is a coarse positioning result of the fault location and needs to be further refined by combining the reflection coefficient.

[0048] This dynamic positioning strategy effectively addresses the shortcomings of traditional positioning methods in complex distribution networks, namely low efficiency and poor accuracy, through branch number adaptation and multi-technology collaboration. For simple distribution networks with few branches, a simplified graph theory method combined with Dijkstra's algorithm can quickly calculate the shortest path, avoiding redundant calculations in complex algorithms and significantly improving positioning efficiency. For complex distribution networks with many branches, the multi-end traveling wave information fusion method, by constructing a time feature matrix and least squares fitting, can effectively eliminate time deviations caused by electromagnetic interference, solving the problem of positioning ambiguity in multi-branch single-end traveling wave methods. Furthermore, the adjustment of the reflection coefficient for line sections further utilizes the wave reflection characteristics of fault points, compensating for the error of positioning based solely on the time matrix and improving the accuracy of section judgment. The dynamic switching between the two strategies achieves a balance between high efficiency in simple scenarios and high accuracy in complex scenarios. At the same time, the deep collaboration between traveling wave characteristics and distribution network topology significantly enhances the anti-interference capability and adaptability of positioning, providing reliable location information for subsequent fault handling.

[0049] Accordingly, when adjusting the line section by combining the reflection coefficient in the traveling wave characteristics, and determining the fault location coordinates of the distribution cable based on the adjusted line section, the specific steps of the implementation plan may include: subdividing the line section where the fault is located into several sub-segments according to a preset length, and calculating the reflection coefficient value of each sub-segment; selecting sub-segments whose corresponding reflection coefficient values ​​are greater than a preset coefficient threshold as candidate fault sub-sections; extracting the time difference between the sensor at the beginning node and the sensor at the end node of the candidate fault sub-section regarding the arrival time of the traveling wave front; calculating the straight-line distance from the fault point to the beginning node based on the traveling wave propagation speed and the time difference, combined with the length of the candidate fault sub-section; and deriving the fault location coordinates of the distribution cable based on the coordinates of the beginning node and the line direction in the distribution network topology information, according to the straight-line distance.

[0050] The faulty line section is a preliminary segment of the line that may be faulty, determined through multi-terminal traveling wave information fusion and least squares fitting. It forms the basis for subsequent refined positioning and is typically coarse (e.g., within 100 meters). The preset length is the sub-segment length (e.g., 10 meters) set when subdividing the faulty line section. The length selection must balance positioning accuracy and computational efficiency; smaller lengths offer higher accuracy but require more computation. The reflection coefficient is a quantitative parameter reflecting the difference in wave impedance between a sub-segment and adjacent lines. Fault points can cause abrupt changes in wave impedance due to insulation failure, resulting in a significantly higher reflection coefficient value for the corresponding sub-segment compared to normal lines. The preset coefficient threshold is the critical value for screening candidate fault sub-segments (e.g., 0.5). Sub-segments exceeding this value are considered potentially faulty areas. This threshold needs to be calibrated based on cable type and operating environment. The selected sub-interval is a segment with a reflection coefficient value greater than a preset threshold selected from the subdivided segments. It represents the precise search range of the fault point (e.g., within ten meters). The time difference is the difference between the time it takes for the traveling wave front to reach the sensor at the first node and the sensor at the last node of the fault candidate sub-interval. It directly relates to the location of the fault point within the candidate sub-interval. The traveling wave propagation speed is the speed at which the traveling wave propagates in the distribution cable. It is determined by the characteristics of the cable insulation medium and is a key parameter for calculating the distance to the fault point. The straight-line distance is the straight-line distance along the line from the fault point to the first node of the fault candidate sub-interval. It is calculated using the traveling wave propagation speed and the time difference and is the basis for deriving the final coordinates. The line direction is the actual physical laying direction of the distribution cable (e.g., east-west direction, turning angle, etc.), derived from the distribution network topology information, and is used to convert the straight-line distance into precise spatial coordinates.

[0051] By subdividing the line section and combining it with the reflection coefficient to screen candidate sub-sections, and utilizing the high reflection coefficient characteristic caused by the sudden change in wave impedance at the fault point, the fault search range can be accurately narrowed, avoiding the inefficiency and error of large-scale blind searches in traditional positioning. The combined calculation of time difference and propagation speed can directly correlate with the spatiotemporal characteristics of traveling wave propagation, providing a more quantitative basis than relying solely on line length estimation. Furthermore, by combining the coordinates of the first-end node with the line route to derive the final coordinates, the matching of the positioning results with the actual physical layout of the distribution network can be ensured, avoiding positioning errors caused by deviations between the straight-line distance and the actual line route. The overall process is refined layer by layer, making full use of the traveling wave reflection characteristics and distribution network topology information, which can significantly reduce the impact of electromagnetic interference, line branches, and other factors on positioning accuracy, keeping the fault positioning error within a smaller range (such as at the meter level), and providing precise location guidance for distribution network operation and maintenance.

[0052] Step 230: After bandpass filtering the ultrasonic partial discharge signal, the maximum amplitude, average amplitude, and repetition frequency of the ultrasonic pulse per unit time are statistically analyzed as partial discharge features. The environmental parameter signal is processed by a 24-hour sliding window, and the maximum, minimum, and trend vectors of temperature and humidity within each window are calculated as environmental features. The traveling wave features, partial discharge features, and environmental features are spliced ​​together according to dimensions to obtain a fused feature set.

[0053] For the embodiments of this disclosure, the ultrasonic partial discharge signal is first subjected to bandpass filtering (preserving the specific frequency band signal corresponding to the partial discharge and filtering out environmental noise), and then the maximum amplitude (upper limit of pulse intensity), average amplitude (average pulse intensity), and repetition frequency (number of pulses per unit time) of the ultrasonic pulse are statistically analyzed. These three quantitative indicators are determined as partial discharge characteristics. Next, for the environmental parameter signal, a 24-hour sliding window processing is used (with a fixed window length of 24 hours, the signal is covered segment by segment according to a preset step size). The maximum and minimum values ​​(extreme environmental conditions) of temperature and humidity and the trend vector of change (quantifying the direction and amplitude of temperature and humidity changes) are calculated in each window. These indicators are determined as environmental characteristics. Finally, the previously extracted traveling wave features, the newly obtained partial discharge features, and the environmental features are sequentially spliced ​​according to the feature dimensions and integrated into a fusion feature set that covers multiple dimensions of fault attributes.

[0054] By bandpass filtering the ultrasonic signal, noise such as mechanical vibration and electromagnetic interference can be filtered out, ensuring that the extracted partial discharge features accurately reflect the state of insulation defects. A 24-hour sliding window processing of environmental parameters preserves the temporal dynamics of the environment while quantifying key information through extreme values ​​and trend vectors, reducing redundant data. Furthermore, the fusion of traveling wave, partial discharge, and environmental features enables a comprehensive characterization of fault transient propagation, local insulation defects, and the influence of the external environment. For example, the traveling wave features of lightning strikes and partial discharge faults may be similar, but the former's environmental features are often accompanied by strong electromagnetic field changes, while the latter's partial discharge features have a more stable pulse frequency. Through multi-feature collaboration, they can be accurately distinguished, significantly improving the classification accuracy of subsequent models for similar faults. Simultaneously, the dimensional stitching integration method ensures the integrity and usability of the feature information.

[0055] Step 240: Optimize the model parameters of the fault classification model using the improved whale algorithm, determine the optimal parameter combination, classify the fused feature set using the fault classification model with the optimal parameter combination, and output the fault type information of the power distribution cable. The improved whale algorithm balances the global search and local optimization capabilities by adjusting the convergence strategy.

[0056] Among them, model parameters are the key adjustable parameters in the fault classification model. Different models have different parameter types (such as the learning step size, number of leaf nodes, regularization coefficient, etc. in LightGBM), which directly affect the model's fitting ability, generalization ability, and classification accuracy. They are the optimization targets for improving the whale algorithm. The optimal parameter combination is the set of model parameter values ​​that achieves the best classification performance of the fault classification model on the fused feature set after optimization by the improved whale algorithm. It is the core achievement of parameter optimization and determines the final classification effect of the model. The convergence strategy is the core mechanism for regulating the parameter optimization process in the improved whale algorithm. By dynamically adjusting the control parameters (such as decreasing them twice with the number of iterations), a balance is achieved between global search in the early stage of iteration (exploring a wide parameter space) and local optimization in the later stage (fine-tuning high-quality parameters). It is the core point of the algorithm improvement.

[0057] In this embodiment of the disclosure, the improved whale algorithm can first be used to optimize the key model parameters (such as learning step size, regularization coefficient, model complexity, etc.) of the fault classification model (such as LightGBM, neural networks, etc.). This algorithm balances the capabilities of global search (extensive exploration of the parameter space in the early stages of iteration) and local optimization (fine-tuning parameters in the later stages of iteration) by adjusting the convergence strategy (such as controlling the parameters to decrease quadratically with the number of iterations), thus selecting the parameter combination that optimizes the model's classification performance (i.e., the optimal parameter combination). Subsequently, this optimal parameter combination is configured into the fault classification model, using the fused feature set as model input. The model learns the mapping relationship between features and fault types to perform classification operations, ultimately outputting specific fault type information of the distribution cable (such as partial discharge, lightning strike, insulation aging, etc.).

[0058] The improved whale algorithm, by balancing global and local optimization capabilities, avoids the inefficiencies and susceptibility to local optima inherent in traditional parameter optimization methods (such as grid search). It can quickly find the optimal parameter combination that fits the fused feature set, fundamentally improving the model's fitting and generalization capabilities. The model with optimal parameters can fully exploit the complementary information of multi-dimensional features in the fused feature set (such as using environmental features to distinguish between lightning strikes and short-circuit faults with similar traveling wave characteristics), significantly improving the accuracy of identifying similar faults. At the same time, the efficient connection between algorithm optimization and model classification ensures the timeliness and reliability of fault type output, providing accurate type basis for subsequent association of fault locations and generation of operation and maintenance plans, effectively supporting the efficiency and accuracy of complex distribution network fault handling.

[0059] Step 250: Associate the fault location coordinates with the fault type information to generate a fault handling plan that includes the location results, fault cause, and maintenance priority.

[0060] For the embodiments of this disclosure, step 250 may specifically include the following steps:

[0061] Step 250-1: Based on the node types in the distribution network topology information and combined with the common distribution patterns of power distribution cable faults, mark the maintenance priority of the corresponding fault types at different locations.

[0062] For the embodiments of this disclosure, firstly, the node types (such as trunk nodes, branch nodes, end-user nodes, hub nodes, etc.) in the distribution network topology information can be extracted to clarify the functional positioning and influence range of different nodes in the distribution network; at the same time, the common distribution patterns of distribution cable faults are combined (i.e., the occurrence frequency and influence of different locations and different fault types are summarized based on historical operation and maintenance data, such as the high occurrence rate of phase-to-phase short circuits and the wide influence range of trunk nodes, while the high occurrence rate of partial discharges of branch nodes has a limited impact); finally, the operation and maintenance priority is marked according to the combination of node type and fault type. For example, phase-to-phase short circuits of trunk hub nodes are marked as the highest priority, and minor partial discharges of branch end nodes are marked as low priority.

[0063] Defining the impact range of a fault location based on node type (e.g., a fault in a trunk node affects a large number of users, while a fault in a branch node only affects a local area) can avoid misjudgments that ignore the importance of location when only looking at the fault type; combining fault distribution patterns can take into account both historical frequency and actual impact, avoiding the subjective arbitrariness of priority labeling.

[0064] Step 250-2: Determine the type of node where the fault is located based on the fault location coordinates, and deduce the cause of the fault based on the fault type information.

[0065] In this embodiment of the disclosure, the specific node to which the fault belongs can first be located in the distribution network topology information based on the determined fault location coordinates, and then the type of the node (such as trunk node, hub node, branch terminal node, etc.) can be determined. Then, combined with the previously identified fault type information, the functional characteristics of the node type (such as hub nodes with many line intersections and complex wave impedance, and branch nodes with easy insulation aging), the node operating environment (such as whether it is in a high-incidence area of ​​lightning strikes or a humid environment), and the typical cause of the fault type (such as partial discharge is often related to insulation defects, and lightning strikes are often related to the external environment and insufficient lightning protection measures of the node), the root cause of the fault can be finally deduced.

[0066] Deducing the cause solely based on the fault type can easily overlook differences in location characteristics. However, by combining the node type determined by the fault location coordinates, the compatibility between the node's structural characteristics, operating conditions, and fault type can be accurately correlated, making the cause deduction more targeted. This precise deduction helps maintenance personnel avoid superficial solutions (such as simply replacing the insulation layer without optimizing the impedance matching of the hub node), addressing the fault problem at its root and significantly reducing the fault recurrence rate.

[0067] Step 250-3: Integrate the fault location coordinates, fault type information, fault cause and maintenance priority, generate a standardized fault handling report, and push it to the distribution network maintenance terminal.

[0068] In this embodiment of the disclosure, the four core information categories—fault location coordinates, fault type information, fault cause, and maintenance priority—acquired in the early stages can be systematically integrated to ensure the completeness and relevance of the information. Then, according to a preset industry or enterprise standard format, the integrated information is compiled into a standardized fault handling report. The report typically includes basic fault information, cause analysis, priority marking, and targeted handling suggestions (such as replacing the insulation layer or optimizing impedance matching). Finally, the report is pushed in real-time to the distribution network maintenance terminal (such as the handheld terminal of maintenance personnel or the monitoring center server) via a communication network, providing direct basis for frontline maintenance and management decisions.

[0069] The integration of these four core information categories avoids the fragmentation of information caused by scattered records of location, type, and cause in traditional operations and maintenance (O&M), ensuring that O&M personnel have complete decision-making basis. Standardized reports can unify output format and content elements, eliminating communication costs caused by differences in records made by different personnel, while also facilitating subsequent archiving and analysis of fault data. Real-time push to O&M terminals enables rapid response with immediate fault discovery, immediate report generation, and immediate instruction issuance, significantly shortening the time lag from fault detection to on-site handling. In addition, the cause analysis and handling suggestions in the report can directly guide O&M personnel to perform precise operations, avoid blind troubleshooting, and rationally allocate resources based on priority marking, ultimately improving fault handling efficiency and reducing recurrence rate, while providing data support for the digital and refined management of distribution network O&M.

[0070] In summary, the technical solution in this application, by collecting multi-mode fault signals composed of high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals, and simultaneously acquiring distribution network topology information such as line length, number of branches, node location, and wave impedance parameters, can overcome the limitations of traditional fault handling, which suffers from a single signal dimension and lack of topology information. This provides multi-dimensional and accurate basic data support for subsequent location and classification. Secondly, after extracting high-frequency traveling wave features, the fault location strategy is dynamically determined based on the number of line branches in the distribution network topology, and the location is performed in conjunction with the traveling wave features and topology information. This approach not only adapts differentiated algorithms to different branch number scenarios to reduce complexity, but also improves the location accuracy of fault coordinates through the combination of features and topology. In terms of efficiency, it can solve the problem of the dramatic increase in complexity of traditional graph theory-based localization algorithms in complex distribution networks. Furthermore, after extracting partial discharge features and environmental features, it fuses traveling wave features to form a multi-dimensional fusion feature set. Combined with an improved whale algorithm-optimized fault classification model for analysis, through multi-feature complementarity and model parameter optimization, it can effectively improve the classification accuracy of similar faults such as partial discharge and lightning strikes, making up for the shortcomings of traditional machine learning in identification. Finally, by associating fault location coordinates with fault type information, it generates a processing solution that includes location results, fault causes, and maintenance priorities. This can break down the barriers between traditional fault location and identification, forming a complete collaborative mechanism for data collection, location, classification, and maintenance guidance, which can fully meet the needs of efficient operation and maintenance of complex distribution networks.

[0071] Furthermore, as Figure 1 and Figure 2 The specific implementation of the method shown in this embodiment provides a power distribution cable fault handling device, such as... Figure 3 As shown, the device includes: a data acquisition module 31, a positioning module 32, a classification module 33, and a generation module 34.

[0072] The acquisition module 31 can be used to acquire multi-mode fault signals of power distribution cables and simultaneously acquire distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location and wave impedance parameters of each line segment.

[0073] The positioning module 32 can be used to extract features from high-frequency traveling wave signals to obtain traveling wave features, determine the fault location strategy based on the number of line branches in the distribution network topology information, and coordinate the traveling wave features with the distribution network topology information to locate the fault location coordinates of the distribution cable.

[0074] The classification module 33 can be used to extract features from ultrasonic partial discharge signals and environmental parameter signals respectively, to obtain partial discharge features and environmental features. The traveling wave features, partial discharge features and environmental features are fused to form a fused feature set. The fused feature set is analyzed based on the fault classification model optimized by the improved whale algorithm to classify similar faults and obtain fault type information of the power distribution cable.

[0075] The generation module 34 can be used to associate the fault location coordinates with the fault type information to generate a fault handling plan that includes the location results, fault causes and maintenance priorities.

[0076] In some embodiments of this application, when extracting features from a high-frequency traveling wave signal to obtain traveling wave features, the positioning module 32 can specifically be used to select the db6 wavelet basis as the decomposition basis function, perform an 8-level discrete wavelet transform on the high-frequency traveling wave signal to obtain the high-frequency detail coefficients of the 1st to 8th levels, calculate the wavelet energy spectrum of each level based on the high-frequency detail coefficients, calculate the traveling wave reflection coefficient based on the wave impedance parameters of each line segment in the distribution network topology information, and determine the wavelet energy spectrum and traveling wave reflection coefficient as traveling wave features.

[0077] In some embodiments of this application, when determining the fault location strategy based on the number of line branches in the distribution network topology information, and performing fault location by coordinating traveling wave characteristics with the distribution network topology information to obtain the fault location coordinates of the distribution cable, the location module 32 can be specifically used as follows: if the number of line branches is less than or equal to a preset threshold, a simplified graph theory location method is adopted, constructing a distribution network topology node graph with node positions as vertices and line lengths as edges, marking the first arrival time of the traveling wave front collected by the high-frequency traveling wave sensor at each node, and calculating the fault location coordinates using Dijkstra's shortest path algorithm; if the number of line branches is greater than the preset threshold, a multi-terminal traveling wave information fusion location method is adopted, based on the distribution network topology information. In the network topology information, the line segments are segmented, and the arrival time of the traveling wave front (TWF) collected by high-frequency traveling wave sensors at the beginning and end nodes of each segment is extracted. Using the line segment number as the row vector and the TWF arrival time as the column vector, time feature matrices are established at the beginning and end of the line segment, respectively. The time feature matrices at the beginning and end of the line segment are then fitted with the fault time matrix using least squares fitting. By minimizing the sum of squared errors between the actual arrival time and the theoretical arrival time, the TWF arrival time deviation caused by electromagnetic interference is eliminated, and the line segment where the fault is located is determined. The line segment is then adjusted based on the reflection coefficient in the traveling wave features, and the fault location coordinates of the distribution cable are determined based on the adjusted line segment.

[0078] In some embodiments of this application, when adjusting the line interval by combining the reflection coefficient in the traveling wave characteristics, and determining the fault location coordinates of the distribution cable based on the adjusted line interval, the positioning module 32 can specifically be used to subdivide the line interval where the fault is located into several sub-segments according to a preset length, calculate the reflection coefficient value of each sub-segment; select sub-segments whose corresponding reflection coefficient values ​​are greater than a preset coefficient threshold as candidate fault sub-intervals; extract the time difference between the sensor at the beginning node and the sensor at the end node of the candidate fault sub-interval regarding the arrival time of the traveling wave front; calculate the straight-line distance between the fault point and the beginning node based on the traveling wave propagation speed and the time difference, combined with the length of the candidate fault sub-interval; and derive the fault location coordinates of the distribution cable according to the straight-line distance based on the coordinates of the beginning node and the line direction in the distribution network topology information.

[0079] In some embodiments of this application, when feature extraction is performed on the ultrasonic partial discharge signal and environmental parameter signal to obtain partial discharge features and environmental features, and the traveling wave features, partial discharge features and environmental features are fused to form a fused feature set, the classification module 33 can be specifically used to perform bandpass filtering on the ultrasonic partial discharge signal, and then statistically analyze the maximum amplitude, average amplitude and repetition frequency of the ultrasonic pulse per unit time as partial discharge features; perform 24-hour sliding window processing on the environmental parameter signal, and calculate the maximum value, minimum value and trend vector of temperature and humidity in each window as environmental features; and concatenate the traveling wave features, partial discharge features and environmental features according to dimensions to obtain the fused feature set.

[0080] In some embodiments of this application, when analyzing the fused feature set based on the fault classification model optimized by the improved whale algorithm, classifying similar faults, and obtaining fault type information of the power distribution cable, the classification module 33 can be specifically used to optimize the model parameters of the fault classification model using the improved whale algorithm, determine the optimal parameter combination, and the improved whale algorithm balances the global search and local optimization capabilities by adjusting the convergence strategy; the fault classification model with the optimal parameter combination is used to classify the fused feature set and output the fault type information of the power distribution cable.

[0081] In some embodiments of this application, the generation module 34 can be specifically used to mark the maintenance priority of different fault types based on the node types in the distribution network topology information and the common distribution patterns of power distribution cable faults; determine the node type where the fault is located based on the fault location coordinates, and deduce the fault cause based on the fault type information; integrate the fault location coordinates, fault type information, fault cause and maintenance priority to generate a standardized fault handling report and push it to the distribution network maintenance terminal.

[0082] It should be noted that other corresponding descriptions of the functional units involved in the power distribution cable fault handling device provided in this embodiment can be found in [reference needed]. Figure 1 and Figure 2 The corresponding descriptions in [the document] will not be repeated here.

[0083] Based on the above, Figure 1 and Figure 2 Accordingly, this embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the above-described method. Figure 1 and Figure 2 The method for handling power distribution cable faults is shown.

[0084] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause an electronic device (such as a personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0085] Based on the above, Figure 1 and Figure 2 The method shown, and Figure 3 To achieve the above objectives, the present application also provides an electronic device, specifically a personal computer, tablet computer, server, or other network device, as shown in the virtual device embodiment. This device includes a storage medium and a processor; the storage medium stores a computer program; the processor executes the computer program to achieve the above-described objectives. Figure 1 and Figure 2 The method for handling power distribution cable faults is shown.

[0086] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0087] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0088] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0089] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platform, or it can be implemented by hardware.

[0090] This invention, through the acquisition of multimodal fault signals comprised of high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals, and simultaneously obtaining distribution network topology information such as line length and branch number, overcomes the limitations of traditional fault handling methods that rely on a single signal dimension and insufficient utilization of topology information. This provides multidimensional and precise data support for subsequent location and classification. Secondly, after extracting high-frequency traveling wave features, the fault location strategy is dynamically determined based on the number of line branches in the distribution network topology. The traveling wave features and topology information are then used in conjunction for location analysis. This approach adapts differentiated algorithms to different branch number scenarios to reduce complexity, and the combination of features and topology improves the accuracy and efficiency of fault location coordinate positioning, thus solving complex fault problems. The traditional graph theory-based localization algorithm for power distribution networks suffers from a significant increase in complexity. Furthermore, by extracting partial discharge and environmental features and fusing traveling wave features to form a multi-dimensional fusion feature set, and combining this with an improved whale algorithm-optimized fault classification model, the classification accuracy for similar faults such as partial discharge and lightning strikes can be effectively improved through multi-feature complementarity and model parameter optimization, thus overcoming the shortcomings of traditional machine learning in identification. Finally, by associating fault location coordinates with fault type information, a processing solution including location results, fault causes, and maintenance priorities is generated. This breaks down the traditional barriers between fault location and identification, forming a complete collaborative mechanism for data collection, location, classification, and maintenance guidance, which can fully meet the needs of efficient operation and maintenance in complex power distribution networks.

[0091] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0092] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A method for handling power distribution cable faults, characterized in that, include: The system collects multi-mode fault signals from power distribution cables and simultaneously acquires distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals, and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location, and wave impedance parameters of each line segment. Feature extraction is performed on the high-frequency traveling wave signal to obtain traveling wave features. A fault location strategy is determined based on the number of line branches in the distribution network topology information. The traveling wave features and the distribution network topology information are used to locate the fault and obtain the fault location coordinates of the distribution cable. Feature extraction is performed on the ultrasonic partial discharge signal and the environmental parameter signal to obtain partial discharge features and environmental features. The traveling wave features, the partial discharge features and the environmental features are fused to form a fused feature set. The fused feature set is analyzed based on the fault classification model optimized by the improved whale algorithm to classify similar faults and obtain the fault type information of the power distribution cable. By associating the fault location coordinates with the fault type information, a fault handling plan is generated that includes the location result, the cause of the fault, and the maintenance priority.

2. The method according to claim 1, characterized in that, The step of extracting features from the high-frequency traveling wave signal to obtain traveling wave features includes: The db6 wavelet basis is selected as the decomposition basis function. The high-frequency traveling wave signal is subjected to 8-level discrete wavelet transform to obtain the high-frequency detail coefficients of the first to eighth levels. The wavelet energy spectrum of each level is calculated based on the high-frequency detail coefficients. Calculate the traveling wave reflection coefficient based on the wave impedance parameters of each line segment in the distribution network topology information; The wavelet energy spectrum and the traveling wave reflection coefficient are used to determine the traveling wave characteristics.

3. The method according to claim 1, characterized in that, The step of determining the fault location strategy based on the number of line branches in the distribution network topology information, and coordinating the traveling wave characteristics with the distribution network topology information to perform fault location and obtain the fault location coordinates of the distribution cable, includes: If the number of line branches is less than or equal to a preset threshold, a simplified graph theory positioning method is adopted. The distribution network topology node graph is constructed with the node position as the vertex and the line length as the edge. The first arrival time of the traveling wave front collected by the high-frequency traveling wave sensor at each node is marked, and the fault location coordinates are calculated by Dijkstra's shortest path algorithm. If the number of line branches is greater than the preset threshold, a multi-terminal traveling wave information fusion positioning method is adopted. Based on the line segments in the distribution network topology information, the arrival time of the traveling wave front collected by the high-frequency traveling wave sensor at the beginning and end nodes of each line segment is extracted. The line segment number is used as the row vector and the arrival time of the traveling wave front is used as the column vector to establish the beginning time feature matrix and the end time feature matrix respectively. The first-end time feature matrix, the last-end time feature matrix, and the fault time matrix are fitted using least squares. By minimizing the sum of squared errors between the actual arrival time and the theoretical arrival time, the travel wave arrival time deviation caused by electromagnetic interference is eliminated, and the line section where the fault is located is determined. The line section is adjusted by combining the reflection coefficient in the travel wave feature, and the fault location coordinates of the power distribution cable are determined based on the adjusted line section.

4. The method according to claim 3, characterized in that, The step of adjusting the line section based on the reflection coefficient in the traveling wave characteristics, and determining the fault location coordinates of the distribution cable based on the adjusted line section, includes: The faulty section of the line is divided into several sub-segments according to a preset length, and the reflection coefficient value of each sub-segment is calculated. The sub-segments whose reflection coefficient values ​​are greater than a preset coefficient threshold are selected as candidate fault sub-intervals; Extract the time difference between the arrival time of the traveling wave front of the sensor at the first end node and the sensor at the last end node of the fault candidate sub-interval; Based on the travel wave propagation speed and the time difference, combined with the length of the candidate fault sub-interval, the straight-line distance from the fault point to the first node is calculated. Based on the coordinates of the first node and the route of the line in the distribution network topology information, the fault location coordinates of the distribution cable are derived according to the straight-line distance.

5. The method according to claim 1, characterized in that, The process involves extracting features from the ultrasonic partial discharge signal and the environmental parameter signal to obtain partial discharge features and environmental features, and then fusing the traveling wave features, the partial discharge features, and the environmental features to form a fused feature set, including: After bandpass filtering the ultrasonic partial discharge signal, the maximum amplitude, average amplitude, and repetition frequency of the ultrasonic pulse per unit time are statistically analyzed and used as partial discharge characteristics. The environmental parameter signals are processed using a 24-hour sliding window method, and the maximum, minimum, and trend vectors of temperature and humidity within each window are calculated as environmental features. The traveling wave feature, the partial discharge feature, and the environmental feature are spliced ​​together according to their dimensions to obtain a fused feature set.

6. The method according to claim 1, characterized in that, The fault classification model based on the improved whale algorithm analyzes the fused feature set, classifies similar faults, and obtains the fault type information of the power distribution cable, including: An improved whale algorithm is used to optimize the model parameters of the fault classification model and determine the optimal parameter combination. The improved whale algorithm balances the global search and local optimization capabilities by adjusting the convergence strategy. The fused feature set is classified using a fault classification model with the optimal parameter combination configured, and the fault type information of the power distribution cable is output.

7. The method according to claim 1, characterized in that, The process of associating the fault location coordinates with the fault type information to generate a fault handling plan that includes the location result, fault cause, and maintenance priority includes: Based on the node types in the power distribution network topology information and combined with the common distribution patterns of power distribution cable faults, the maintenance priorities corresponding to different fault types at different locations are marked. The type of node where the fault is located is determined based on the fault location coordinates, and the cause of the fault is deduced by combining the fault type information. The fault location coordinates, fault type information, fault cause, and maintenance priority are integrated to generate a standardized fault handling report, which is then pushed to the distribution network maintenance terminal.

8. A power distribution cable fault handling device, characterized in that, include: The acquisition module is used to acquire multi-mode fault signals of power distribution cables and simultaneously obtain distribution network topology information. The multi-mode fault signals include at least high-frequency traveling wave signals, ultrasonic partial discharge signals and environmental parameter signals. The distribution network topology information includes at least line length, number of line branches, node location and wave impedance parameters of each line segment. The positioning module is used to extract features from the high-frequency traveling wave signal to obtain traveling wave features, determine the fault location strategy based on the number of line branches in the distribution network topology information, and coordinate the traveling wave features with the distribution network topology information to perform fault location and obtain the fault location coordinates of the distribution cable. The classification module is used to extract features from the ultrasonic partial discharge signal and the environmental parameter signal respectively to obtain partial discharge features and environmental features. The traveling wave features, partial discharge features and environmental features are fused to form a fused feature set. The fused feature set is analyzed based on the fault classification model optimized by the improved whale algorithm to classify similar faults and obtain the fault type information of the power distribution cable. The generation module is used to associate the fault location coordinates with the fault type information to generate a fault handling plan that includes the location result, fault cause and maintenance priority.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images