Power transmission line fault backtracking analysis method and system and storage medium thereof

Abstract

The invention provides a power transmission line fault backtracking analysis method and system and a storage medium, and the method comprises the following steps: carrying out the hierarchical processing of lightning events, traveling wave waveforms, meteorological parameters and video image data of a power transmission line, and obtaining corresponding fault characterization features; and carrying out logic combination on the multi-source fault characterization characteristics according to preset judgment conditions of all fault types, obtaining the fault types and determining fault positions and time. The precision, efficiency and reliability of power transmission line fault diagnosis are effectively improved, and the fault type, the fault occurrence position and the fault reason are accurately determined.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system fault diagnosis, and in particular relates to a transmission line fault backtracking analysis method, system and storage medium thereof. Background Art

[0002] Transmission lines operate in a complex environment. Lightning strikes, external damage, and extreme weather conditions often cause transmission line failures, severely impacting the safe and stable operation of the power grid. Current transmission line fault diagnosis methods often rely on single monitoring methods. For example, traveling wave monitoring methods can only capture the waveform characteristics and arrival time of the fault, making it difficult to pinpoint the cause of the fault. Lightning location systems can monitor the location and intensity of lightning strikes but cannot effectively identify non-lightning-induced faults. Meteorological monitoring methods primarily provide information on regional meteorological conditions such as wind speed and rainfall, but cannot accurately determine the specific fault type and location. While video surveillance technology can provide a visual representation of the fault site, it struggles to identify the cause and accurately locate the fault.

[0003] Since the above-mentioned single technical means lack comprehensive data support during fault diagnosis, it is difficult to accurately identify the fault type and precisely locate the fault when faced with complex multi-source fault situations. It is also unable to effectively deal with the high uncertainty of the fault causes in complex environments, resulting in insufficient fault diagnosis accuracy and low efficiency, making it difficult to quickly formulate maintenance strategies. Summary of the Invention

[0004] The purpose of the present invention is to address the shortcomings of the above-mentioned background technology and provide a transmission line fault retrospective analysis method, system and storage medium thereof, and a method for integrated analysis of multiple data sources such as lightning events, traveling wave waveforms, meteorological parameters and video images to improve the accuracy, efficiency and reliability of transmission line fault diagnosis and accurately determine the fault type, location and cause.

[0005] The technical solution adopted by the present invention is: a transmission line fault retrospective analysis method, comprising the following steps:

[0006] Perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features;

[0007] The multi-source fault characterization features are logically combined according to the preset judgment conditions of each fault type to obtain the fault type and determine the fault location and time.

[0008] In the above technical solution, before the layered processing, the following is also included:

[0009] Lightning events, traveling wave waveforms, meteorological parameters and video image data are annotated with unified timestamps and geographic location labels, and the annotated multi-source data are aggregated into several event clusters based on the preset time window and spatial range.

[0010] In the above technical solution, the layered processing includes:

[0011] Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data;

[0012] For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information;

[0013] Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics;

[0014] Identify key frame anomalies in video image data and extract fault scene types.

[0015] In the above technical solution, the multi-condition fusion determination process includes:

[0016] When all of the following conditions are met, it is determined to be a lightning-type fault:

[0017] The traveling wave waveform type is single peak or multi-peak;

[0018] A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

[0019] In the above technical solution, the multi-condition fusion determination process includes:

[0020] When all of the following conditions are met, it is determined to be a meteorological fault:

[0021] The traveling wave waveform type is an oscillating waveform or an irregular waveform;

[0022] The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window;

[0023] The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold;

[0024] Rain or strong winds were detected in the video keyframes.

[0025] In the above technical solution, the multi-condition fusion determination process includes:

[0026] When all of the following conditions are met, it is determined to be an external force fault:

[0027] The traveling wave waveform type is a single peak waveform or an irregular waveform;

[0028] The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold;

[0029] Meteorological parameters do not meet the preset wind speed and rainfall thresholds;

[0030] External force damage was detected in the video key frames.

[0031] In the above technical solution, the multi-condition fusion determination process further includes:

[0032] When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

[0033] The above technical solution also includes generating a fault backtracking analysis report based on the fault type, fault location and time.

[0034] The above technical solution also includes the following steps: according to the fault type determined by the multi-condition fusion judgment step, based on the mapping relationship between the multi-source fault characterization features used and the preset causes, the corresponding fault cause label is automatically generated; and the fault cause label, the characterization features that trigger the fault cause label and their values ​​are included in the fault backtracking analysis report.

[0035] In the above technical solution, after the fault backtracking analysis report is generated, the following visual display steps are also included:

[0036] Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols;

[0037] Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window;

[0038] At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence;

[0039] When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

[0040] The present invention also provides a transmission line fault tracing and analysis system, comprising:

[0041] Multi-source data analysis module, used to perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features;

[0042] The fault judgment module is used to logically combine the multi-source fault characterization features according to the preset judgment conditions of each fault type, obtain the fault type and determine the fault location and time.

[0043] The above technical solution also includes a fault backtracing module, which is used to generate a fault backtracing analysis report based on the fault type, fault location and time.

[0044] The above technical solution also includes a data acquisition and fusion platform, which is used to mark lightning events, traveling wave waveforms, meteorological parameters and video image data with unified timestamps and geographic locations, and aggregate the marked multi-source data into several event clusters based on preset time windows and spatial ranges and send them to the multi-source data analysis module.

[0045] In the above technical solution, the hierarchical processing process of the multi-source data analysis module includes:

[0046] Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data;

[0047] For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information;

[0048] Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics;

[0049] Identify key frame anomalies in video image data and extract fault scene types.

[0050] In the above technical solution, the multi-condition fusion judgment process of the multi-source data analysis module includes:

[0051] When all of the following conditions are met, it is determined to be a lightning-type fault:

[0052] The traveling wave waveform type is single peak or multi-peak;

[0053] A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

[0054] In the above technical solution, the multi-condition fusion judgment process of the multi-source data analysis module includes:

[0055] When all of the following conditions are met, it is determined to be a meteorological fault:

[0056] The traveling wave waveform type is an oscillating waveform or an irregular waveform;

[0057] The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window;

[0058] The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold;

[0059] Rain or strong winds were detected in the video keyframes.

[0060] In the above technical solution, the multi-condition fusion judgment process of the multi-source data analysis module includes:

[0061] When all of the following conditions are met, it is determined to be an external force fault:

[0062] The traveling wave waveform type is a single peak waveform or an irregular waveform;

[0063] The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold;

[0064] Meteorological parameters do not meet the preset wind speed and rainfall thresholds;

[0065] External force damage was detected in the video key frames.

[0066] In the above technical solution, the multi-condition fusion process steps of the multi-source data analysis module also include:

[0067] When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

[0068] In the above technical solution, the fault backtracing module is also used to automatically generate a corresponding fault cause label based on the fault type determined by multi-condition fusion and the mapping relationship between the multi-source fault characterization features used and the preset causes; and include the fault cause label, the characterization features that trigger the fault cause label, and their values ​​in the fault backtracing analysis report.

[0069] The above technical solution also includes a visualization and decision support module, which is used after the fault backtracking analysis report is generated.

[0070] Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols;

[0071] Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window;

[0072] At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence;

[0073] When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

[0074] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the transmission line fault backtracking analysis method described in the above technical solution.

[0075] The present invention also provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the transmission line fault backtracking analysis method described in the above technical solution by executing the computer instructions.

[0076] The beneficial effects of the present invention are: by implementing layered processing on multi-source time series data and extracting fault characterization features, and then combining multi-condition fusion judgment and generating backtracking reports, the present invention establishes a closed-loop fault diagnosis and backtracking process, which can significantly improve the accuracy of fault type identification and positioning accuracy, and at the same time realize end-to-end automation from fault detection to report output, reducing the workload of manual analysis and improving diagnostic efficiency.

[0077] Furthermore, the present invention uniformly timestamps and labels the multi-source data with geographic locations before hierarchical processing, and aggregates them into event clusters, which can ensure the temporal and spatial consistency of each source data, avoid misjudgment due to timing deviation or spatial mismatch, effectively filter out irrelevant noise, and provide a reliable preprocessing basis for subsequent hierarchical verification and feature extraction.

[0078] Furthermore, the present invention ensures that the extracted characterization features are highly correlated with the fault entity by performing refined verification and feature extraction of traveling wave waveforms, lightning events, meteorological parameters and video images for each data source, which not only improves the reliability of single-source data processing, but also provides accurate and traceable multi-dimensional information for fusion judgment.

[0079] Furthermore, the present invention clearly defines the judgment conditions for lightning-type faults (waveform type + time-space matching + intensity threshold), which can accurately identify lightning faults, avoid omissions or false alarms that are easily caused by relying on a single data source (such as only traveling waves or only lightning data), and improve the diagnostic accuracy of lightning faults.

[0080] Furthermore, the four conditions set by the present invention for meteorological faults, namely oscillation waveform, no lightning strike matching, meteorological threshold and video anomaly, can accurately grasp the fault scenarios caused by weather factors such as strong winds or heavy rains, supplement the problem of easy missed diagnosis when relying only on traveling waves or single meteorological data, and enhance the ability to identify meteorological-induced faults.

[0081] Furthermore, the present invention can effectively identify faults caused by external damage such as foreign objects hanging on wires and fallen trees through multi-condition judgment of external force-type faults (single peak / irregular waveform + no lightning strike + normal weather + video external force evidence), filling the blind spots of traditional methods in external force damage scenarios and improving the system's robustness to various external factors.

[0082] Furthermore, when multi-source features simultaneously meet the conditions of multiple fault types, the present invention calls a preset machine learning model or expert system to perform complex fault analysis, which can cope with complex fault scenarios and achieve in-depth analysis of the superposition of multiple fault factors, thereby enhancing the system's adaptability and diagnostic coverage in complex environments.

[0083] Furthermore, the present invention automatically generates a fault cause label and incorporates the characterizing features and values ​​that trigger the label into the report, which not only provides a clear causal link, but also makes the fault report traceable and credible, making it easier for operation and maintenance personnel to quickly understand the diagnostic basis and formulate targeted maintenance plans, reducing the risk of manual secondary judgment.

[0084] Furthermore, the present invention intuitively presents the temporal and spatial evolution of the fault and various analysis results through the visualization steps of map annotation, timeline display, and linkage of multi-source curves and key video frames, enabling operation and maintenance personnel to grasp the overall picture of the fault at a glance, speeding up the response speed of fault site location and maintenance decision-making, and improving overall operation and maintenance efficiency and safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0085] Figure 1 Schematic diagram of the method flow of the present invention;

[0086] Figure 2 It is a schematic diagram of the hierarchical processing logic of the present invention;

[0087] Figure 3 A logical diagram of the visualization process of the present invention;

[0088] Figure 4 Schematic diagram of the system architecture of the present invention. DETAILED DESCRIPTION

[0089] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments to facilitate a clear understanding of the present invention, but they do not constitute a limitation to the present invention.

[0090] like Figure 1 As shown, the present invention provides a transmission line fault retrospective analysis method, comprising the following steps:

[0091] Perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features;

[0092] The multi-source fault characterization features are combined according to the preset conditions of each fault type for logical fusion judgment to obtain the fault type and determine the fault location and time.

[0093] Specifically, the method further includes the following steps: generating a fault backtracking analysis report according to the fault type, fault location and time.

[0094] Specifically, before the layered processing:

[0095] Lightning events, traveling wave waveforms, meteorological parameters, and video image data are annotated with unified timestamps and geographic locations, and the aligned multi-source data are aggregated into several event clusters based on preset time windows and spatial ranges; subsequent hierarchical processing is performed based on each event cluster.

[0096] Preferably, the sources of the multi-source data in this embodiment are as follows:

[0097] Lightning location system: Deploy wide-area lightning location or local lightning location system to monitor the time, location and intensity of lightning strikes.

[0098] Traveling wave monitoring equipment: installed at both ends or in the middle of the line to collect fault traveling wave waveforms and arrival times.

[0099] Meteorological monitoring equipment: including anemometers, rain gauges and weather stations, used to collect regional wind speed, rainfall and other data.

[0100] Video surveillance equipment: deployed in key areas, covering key sections of the line and fault-prone locations.

[0101] Before layered processing, the process of aggregating the above multi-source data into event clusters specifically includes:

[0102] 1. Data preprocessing and annotation

[0103] 1.1 Timestamp Unification

[0104] All monitoring equipment (lightning location systems, traveling wave monitoring equipment, weather stations, and video cameras) should be configured to automatically attach standard UTC timestamps using GPS or NTP. After receiving the raw data, the acquisition module aligns the timestamps of each data type to the millisecond level, with a maximum allowable time deviation of ±100ms.

[0105] 1.2 Geographic Location Annotation

[0106] Lightning location data: It comes with longitude and latitude coordinates, which are converted into the line measurement point coordinate system (such as mileage stake numbers along the line) through GIS services.

[0107] Traveling wave monitoring data: The monitoring device is installed at both ends or in the middle of the line. When collecting data, the GPS positioning coordinates (Lat / Lon) of the device are mapped to the nearest line stake number.

[0108] Meteorological monitoring data: Weather stations are located along and around the route. Each data record contains the geographical coordinates of the weather station, which is also mapped to the route stake number.

[0109] Video surveillance data: The location of fixed / dome cameras is recorded on the installation drawings, and the camera installation location data is displayed when the video frame or key frame is extracted.

[0110] 1.3 Data Quality Filtering

[0111] After receiving the labeled multi-source data, the edge computing node first performs noise removal:

[0112] (1) Automatic filtering of traveling wave peaks that exceed monitoring sensitivity;

[0113] (2) invalid or erroneous GPS coordinates discarded during lightning events;

[0114] (3) If the fault of the weather station sensor exceeds the threshold, the data at that moment will be skipped;

[0115] (4) Key frames with blurred video images and frame rates lower than 15fps are discarded.

[0116] 2. Time-space event cluster aggregation

[0117] 2.1 Preset Aggregation Parameters

[0118] Time window: The default setting is ±500ms, and users can adjust it to ±100ms to ±2s in the platform management interface.

[0119] Spatial range: The default is 5km, and the configurable range is 1km to 10km.

[0120] 2.2 Aggregation Algorithm

[0121] (1) Multi-source time series records sorted by timestamps are preliminarily grouped within a sliding time window;

[0122] (2) For the records within the same time window, calculate the distance between the marked line pile coordinates; if the maximum distance does not exceed the spatial range, the corresponding data are classified as the same event initial cluster;

[0123] (3) The sliding window is continuously used to traverse all time series records, and the DBSCAN algorithm is used to cluster the distance and time dimensions to form the final event cluster list.

[0124] 2.3 Event Cluster Output

[0125] Each event cluster contains a set of lightning, traveling wave, meteorological and video records that are highly correlated in time and space. The event cluster object contains: cluster ID; time range (start / end UTC time); spatial range (start / end line stake number); index list of each source data; aggregation center (average timestamp, average stake number).

[0126] Specifically, the layered processing includes:

[0127] Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data;

[0128] For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information;

[0129] Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics;

[0130] Identify key frame anomalies in video image data and extract fault scene types.

[0131] like Figure 2 As shown, preferably, the following four layers of hierarchical processing are performed on the aligned and aggregated event clusters in sequence, each layer completing both verification (checking) and synchronously extracting characterization features to form a multi-source fault characterization feature set:

[0132] First layer: preliminary analysis of traveling wave waveform

[0133] 1. Waveform preprocessing

[0134] The collected traveling wave monitoring signal is first low-pass filtered below 10kHz to filter out high-frequency noise;

[0135] The waveform is digitized using an A / D converter with a sampling rate of 1MHz and sent to an FPGA or DSP for subsequent processing.

[0136] 2. Waveform type identification

[0137] Use zero-crossing detection and peak detection algorithms to perform statistics on waveform amplitude and extreme points:

[0138] If only a single main peak is detected without subsequent secondary peaks, it is determined to be a "single-peak waveform", which may be caused by instantaneous lightning strikes or short-term external force effects;

[0139] If two or more peaks are detected within the fault time window and the interval between adjacent peaks is less than 1ms, it is determined to be a "multi-peak waveform", which may be caused by continuous lightning strikes or discharges at multiple grounding points;

[0140] If the waveform continues to oscillate and decay after the main peak, it is judged as an "oscillating waveform", which may be caused by equipment insulation flashover or complex external force interference.

[0141] 3. Fault time window extraction

[0142] Taking the rising edge time T0 of the first main peak as the center, it extends forward / backward by ΔT (typical value ±0.2ms) to form the traveling wave fault time window [T0–ΔT, T0+ΔT];

[0143] The time domain features of the waveform, such as peak amplitude, rise time, and pulse width, are recorded together for subsequent fusion judgment.

[0144] Second layer: Lightning location verification

[0145] 1. Time matching verification

[0146] The event time T of the lightning location system i Compared with the endpoint of the traveling wave fault time window, if T i ∈[T0–ΔT,T0+ΔT] is matched by time;

[0147] 2. Spatial distance verification

[0148] Lightning location gives the longitude and latitude (Lati, Loni), and the horizontal distance D to the nearest line point is calculated in the line geographic model (using the Haversine algorithm). Spatial verification is only passed when D ≤ 5 km.

[0149] 3. Intensity threshold verification

[0150] If the lightning intensity I i ≥20kA (or user-set threshold), the strength check is passed;

[0151] At the same time, within the same fault time window, count whether there are two or more lightning strikes with a time interval of less than 1s to identify multi-grounding point discharge phenomena;

[0152] 4. Feature Extraction

[0153] For lightning events that pass the three checks, extract their intensity value I i , longitude and latitude coordinates and the number of multiple grounding points are used as lightning characterization features.

[0154] The third layer: meteorological data verification

[0155] 1. Wind speed verification and extraction

[0156] Obtain wind speed time series data V(t) with a resolution of 1 Hz from the weather station. If V(t) ≥ 25 m / s at any moment within the traveling wave fault time window, mark it as "high wind speed anomaly" and record the peak value.

[0157] 2. Rainfall verification and extraction

[0158] The accumulated rainfall R(t) of the rain gauge is collected synchronously. If ΔR = R(t2) – R(t1) ≥ 30 mm / h in the same time window, it is marked as “heavy rainfall anomaly” and the maximum accumulated value in the window is extracted.

[0159] 3. Composite weather tips

[0160] When the wind speed and rainfall exceed the standard at the same time, the "convective composite fault" sign is recorded for the determination of complex faults.

[0161] Layer 4: Video Surveillance Verification

[0162] 1. Keyframe extraction

[0163] The frame difference method and optical flow method are used to analyze the HD video and extract possible fault key frames at 2fps;

[0164] 2. Abnormal scene identification

[0165] Apply pre-trained object detection models (such as YOLOv5) to keyframes to identify thunderstorm flashes, foreign objects hanging on wires, fallen trees, strong winds and rain, etc.

[0166] 3. Verification and feature extraction

[0167] If any of the above anomalies is detected, the video is verified and the corresponding frame timestamp, frame number and anomaly type label are extracted;

[0168] If no anomaly is detected within the time window, the layer is marked as “no keyframe anomaly”.

[0169] After the above four layers of processing are completed, the system obtains a set of highly correlated and traceable multi-source fault characterization features (including traveling wave time window / type, lightning point coordinates / intensity, multiple grounding point marks, wind speed / rainfall peaks, video anomaly types, etc.), providing accurate and reliable input for subsequent multi-condition fusion judgment.

[0170] Specifically, the multi-condition fusion determination process includes:

[0171] When all of the following conditions are met, it is determined to be a lightning-type fault:

[0172] The traveling wave waveform type is single peak or multi-peak;

[0173] A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

[0174] The multi-condition fusion determination process includes:

[0175] When all of the following conditions are met, it is determined to be a meteorological fault:

[0176] The traveling wave waveform type is an oscillating waveform or an irregular waveform;

[0177] The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window;

[0178] The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold;

[0179] Rain or strong winds were detected in the video keyframes.

[0180] The multi-condition fusion determination process includes:

[0181] When all of the following conditions are met, it is determined to be an external force fault:

[0182] The traveling wave waveform type is a single peak waveform or an irregular waveform;

[0183] The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold;

[0184] Meteorological parameters do not meet the preset wind speed and rainfall thresholds;

[0185] External force damage was detected in the video key frames.

[0186] The multi-condition fusion determination process further includes:

[0187] When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

[0188] According to the fault type determined by the multi-condition fusion judgment step, the corresponding fault cause label is automatically generated based on the mapping relationship between the multi-source fault characterization features used and the preset cause; and the fault cause label, the characterization features that trigger the fault cause label and their values ​​are included in the fault backtracking analysis report.

[0189] Preferably, the process of determining the fault type and generating the fault cause label by integrating multiple conditions in this embodiment can be divided into the following stages:

[0190] 1. Feature aggregation and preprocessing

[0191] The system first aggregates the multi-source characterization features extracted during the hierarchical processing phase—including traveling wave waveform type and fault time window, a list of verified lightning events (time, location, intensity), meteorological peak parameters (wind speed, rainfall), and abnormal scene types in video key frames—into a structured feature object. For lightning events, the system also removes records from memory that do not meet the time window, spatial distance, or intensity thresholds, retaining only valid events that are highly correlated with the fault. Based on this structured feature object, whisker fault type and cause analysis is performed.

[0192] 2. Determination of lightning-type faults

[0193] The system first determines whether the traveling wave waveform type is single-peak or multi-peak; if so, it further checks whether the list of lightning events remaining after preprocessing is not empty (that is, there is at least one lightning strike that meets the time, space and intensity conditions).

[0194] If a "thunderstorm" scene is detected in the video on this basis, the fault will be immediately determined to be a "lightning strike type fault".

[0195] If no thunderstorm is captured in the video, but the lightning event fully meets the matching conditions, the system will still determine it as a "lightning strike type fault."

[0196] 3. Meteorological fault determination

[0197] If the traveling wave waveform type is an oscillating or irregular waveform, and there is no valid lightning event after preprocessing, the system will then judge the meteorological characteristics: as long as any indicator of wind speed or rainfall exceeds the preset threshold (such as wind speed ≥ 25m / s or rainfall ≥ 30mm / h), and "rainfall" or "strong wind" anomalies appear in the video key frame, the system will determine it as a "meteorological fault".

[0198] 4. External force fault determination

[0199] When the traveling wave waveform is a single peak or irregular waveform, and there is no effective lightning event and no meteorological threshold conditions are met, the system finally retrieves the video key frame:

[0200] If external force damage scenarios such as "foreign objects hanging on the wires" or "trees falling" are detected, it will be determined as an "external force type failure".

[0201] 5. Complex fault determination

[0202] When a transmission line fault event meets two or more of the following criteria: lightning strike, meteorological, or external force, the system considers the fault to be a "complex fault" and performs in-depth analysis and composite cause generation according to the following process:

[0203] The system aggregates all the characterization features extracted by the aforementioned four-layer hierarchical processing into a composite feature set in a fixed format, which includes:

[0204] Traveling wave waveform type code and fault time window center;

[0205] The number of lightning events, their average intensity and location after verification of time, space and intensity;

[0206] The maximum wind speed value and cumulative rainfall within the time window;

[0207] All anomaly labels detected in the video keyframes (e.g., “thunderstorm,” “fallen tree”).

[0208] If confidence scores for each layer are calculated in the routine judgment, they will also be added to the feature set to enhance the discriminative power of subsequent analysis.

[0209] During the system deployment phase, machine learning models (such as random forests, gradient boosting trees, or neural networks) are trained using historical fault datasets. Training samples are annotated manually or by an expert system. Each sample contains the aforementioned composite feature set and the corresponding composite fault cause label (e.g., "lightning strike + strong wind"). After training, the model is solidified and imported into the inference engine. A set of expert rule-based inference libraries can also be constructed as a supplement or alternative to the model.

[0210] When the system detects both lightning matching and meteorological anomalies in an event, or cross-prompts of video and traveling wave characteristics, it automatically enters complex analysis mode:

[0211] If the machine learning model has been deployed and is available, the system inputs the composite feature set into the model and outputs the most likely composite fault cause label.

[0212] If the model is temporarily unavailable or the user selects expert rules, the expert system is called to match the corresponding composite scenario rules from the rule base and infer the composite cause.

[0213] Regardless of whether the model or the expert system returns a composite cause, it serves as the fault cause label. The system also tracks and organizes all key features and their values ​​used for composite judgment—for example, lightning strike intensity, peak wind speed, and abnormal scenes in video frames—to form a "chain of evidence" list.

[0214] In the final fault traceback report, the "complex fault" and its combined cause label, such as "lightning strike + strong wind," are first given, followed by a paragraph clearly outlining the basis for the judgment:

[0215] "After an in-depth analysis of the system, we determined that the fault type was complex, primarily due to the combined effects of lightning strikes and strong winds: a 24kA lightning strike occurred at 12:05:03, a peak wind speed of 27m / s was recorded at 12:05:04, and video surveillance captured images of violent swinging of the conductors." This explanation not only helps operation and maintenance personnel understand the interaction of multiple causes, but also provides an accurate basis for subsequent maintenance strategies.

[0216] 6. Generate fault cause label

[0217] During system deployment, a "fault type → cause label" mapping table is first loaded from the configuration file or database. For example, "LightningFault" is mapped to "Lightning Strike," "WeatherFault" is mapped to "High Winds / Heavy Rainfall," and "ExternalForceFault" is mapped to "External Force Damage." "ComplexFault" is reserved for complex causes returned by subsequent models. This mapping table facilitates direct access to human-readable descriptions of the fault cause based on the judgment results.

[0218] After completing the multi-condition fusion judgment, the system will obtain a high-level fault type (such as lightning strike, meteorological type, external force type, or complex type). The system first searches the mapping table for the initial cause label corresponding to this type. If it is determined to be a "complex fault," it will further call a pre-trained machine learning model or rule-based expert system, input all the characteristic features of the fault, and let the model output a composite cause label such as "lightning strike + strong wind" or "external force + rainfall."

[0219] To ensure the traceability of the report, the system will backtrack and collect the key features and their specific values ​​that this judgment relies on:

[0220] For the traveling wave layer, the main peak time point and waveform type were recorded;

[0221] For the lightning layer, the latitude, longitude, and intensity of time-matched lightning events are recorded;

[0222] For the meteorological layer, record the maximum wind speed or accumulated rainfall detected;

[0223] For the video layer, the key frame number and abnormality type of the detected abnormal scene are recorded.

[0224] These "trigger evidences" are organized into a list, indicating the source level, feature name and value of each piece of evidence.

[0225] The system assigns weights based on the importance of different pieces of evidence and normalizes the "matching degree" of each piece of evidence to calculate an overall confidence score. For example, traveling wave evidence might have a weight of 0.4, lightning evidence 0.3, meteorological evidence 0.2, and video evidence 0.1, ultimately resulting in a confidence score between 0 and 1, indicating the reliability of the judgment.

[0226] The system combines the selected reason label with the list of evidence that triggers the conclusion and automatically generates a text description. For example:

[0227] "After fusion of multiple conditions, the fault type is determined to be a 'lightning strike fault'. This is based on the following: single-peak traveling wave waveform (T0 = 2025-04-2012:05:03); lightning strike event (T1 = 12:05:03.1, location 30.50°N / 114.35°E, intensity 24kA); and video keyframe capturing the thunderstorm (frame number 453). We recommend checking the insulation and grounding conditions at the corresponding locations."

[0228] If there is a confidence level, "Judgment confidence: 0.87" will also be displayed at the beginning of the paragraph.

[0229] Ultimately, the system writes the following to the report data structure:

[0230] Fault type and corresponding cause label; name, value and source layer of each piece of evidence that triggers the judgment; optional confidence score.

[0231] The report contains structured fields for easy program processing, as well as human-readable text descriptions, which make it easier for operation and maintenance personnel to quickly understand and take targeted maintenance measures.

[0232] Specifically, after the fault backtracking analysis report is generated, the following visual display steps are also included:

[0233] Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols;

[0234] Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window;

[0235] At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence;

[0236] When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

[0237] like Figure 3As shown, preferably, in this embodiment, after the fault backtracking analysis report is generated, the system also visually presents the multi-source analysis results to the operation and maintenance personnel through visualization. The specific implementation process can be divided into the following steps:

[0238] 1. Geographic map annotation

[0239] The system loads the geographic topology information of the transmission lines (such as GeoJSON data of the line corridors) and displays it as an overlay layer on the map interface:

[0240] Use different icons or symbols (such as lightning, raindrop, and broken wire) to represent fault types such as "lightning strike," "weather," and "external force."

[0241] The fault location coordinates (latitude and longitude) in each traceback report are used as the anchor point of the icon and accurately marked on the corresponding line segment;

[0242] The map supports panning, zooming, and layer switching, allowing operations and maintenance personnel to freely view the distribution of faults in different areas.

[0243] 2. Timeline distribution presentation

[0244] Below or on the side of the map, the system provides a horizontal timeline control to display the temporal distribution of all fault events:

[0245] The timeline uses scales or heat bars to indicate the time point and intensity of each fault event.

[0246] Operations and maintenance personnel can drag the time slider or select a time range. The system will immediately highlight only the fault points within that time window on the map, making it easy to quickly locate the fault cluster area within a specific period of time.

[0247] 3. Detailed drawing of multi-source data

[0248] When the operator clicks a fault point on the map or selects a time window in the timeline, the system will draw the source analysis charts and video key frames of the fault instance in the same interface or adjacent panels in the following order:

[0249] Traveling wave waveform curve: Displays the traveling wave signal waveform collected within the fault time window in the form of a time domain line graph, marking the main peak position and waveform type annotation;

[0250] Lightning strike intensity and spatial distribution map: Plot all lightning location points corresponding to the event on a small map or scatter plot, and use bubble size or color depth to reflect the lightning strike intensity;

[0251] Meteorological time series curve: Use a dual-axis line graph to display the wind speed and cumulative rainfall curves within the time window, and clearly mark the threshold exceeding intervals;

[0252] Video keyframe: Automatically locates the video frame at the moment the fault occurs and highlights the detected abnormal targets (such as lightning, fallen trees or swaying wires) in the picture.

[0253] 4. Interactive linkage refresh

[0254] Click another fault point on the map, and the timeline and various charts will automatically jump to the corresponding period of the event, and update the waveform, weather curve and video frame;

[0255] Drag the slider on the time axis or zoom in or out on the time range. The fault markers on the map will be updated synchronously with the multi-source curve graph, showing only the content within the selected time window.

[0256] Clicking a specific data point on any data panel (such as a waveform chart or weather map) can also trigger the linkage between the map and timeline, highlighting the corresponding fault location and time.

[0257] 5. Layout and performance optimization

[0258] The front end uses a column layout: the interactive map is displayed on the left, and the timeline and multi-source panel are at the bottom or right, ensuring that each view is visible to each other;

[0259] All chart and map annotation data is obtained in real time through asynchronous interfaces, and client-side caching or backend Redis cache is used to accelerate responses.

[0260] The linkage between panels is achieved through an event publish / subscribe mechanism, ensuring that after user operation, each view can be redrawn within milliseconds, providing a smooth interactive experience.

[0261] Through the above visualization steps, operation and maintenance personnel can not only intuitively understand the spatial and temporal distribution of faults, but also quickly view the original waveforms, lightning points, meteorological curves and video evidence supporting the judgment, greatly improving the speed and accuracy of fault location and maintenance decisions.

[0262] Example 2

[0263] like Figure 4 As shown, the present invention also provides a transmission line fault tracing analysis system, comprising:

[0264] Multi-source data analysis module, used to perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features;

[0265] The fault judgment module is used to logically combine the multi-source fault characterization features according to the preset judgment conditions of each fault type, obtain the fault type and determine the fault location and time.

[0266] Specifically, it also includes a fault backtracking module, which is used to generate a fault backtracking analysis report according to the fault type, fault location and time.

[0267] Specifically, it also includes a data acquisition and fusion platform, which is used to mark lightning events, traveling wave waveforms, meteorological parameters and video image data with unified timestamps and geographic locations, and aggregate the marked multi-source data into several event clusters based on preset time windows and spatial ranges and send them to the multi-source data analysis module; the multi-source data analysis module performs subsequent hierarchical processing based on each event cluster.

[0268] Specifically, the hierarchical processing of the multi-source data analysis module includes:

[0269] Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data;

[0270] For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information;

[0271] Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics;

[0272] Identify key frame anomalies in video image data and extract fault scene types.

[0273] Specifically, the multi-condition fusion judgment process of the multi-source data analysis module includes:

[0274] When all of the following conditions are met, it is determined to be a lightning-type fault:

[0275] The traveling wave waveform type is single peak or multi-peak;

[0276] A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

[0277] In the above technical solution, the multi-condition fusion judgment process of the multi-source data analysis module includes:

[0278] When all of the following conditions are met, it is determined to be a meteorological fault:

[0279] The traveling wave waveform type is an oscillating waveform or an irregular waveform;

[0280] The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window;

[0281] The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold;

[0282] Rain or strong winds were detected in the video keyframes.

[0283] When all of the following conditions are met, it is determined to be an external force fault:

[0284] The traveling wave waveform type is a single peak waveform or an irregular waveform;

[0285] The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold;

[0286] Meteorological parameters do not meet the preset wind speed and rainfall thresholds;

[0287] External force damage was detected in the video key frames.

[0288] When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

[0289] Specifically, the fault backtracing module is also used to automatically generate a corresponding fault cause label based on the fault type determined by multi-condition fusion and the mapping relationship between the multi-source fault characterization features used and the preset causes; and include the fault cause label, the characterization features that trigger the fault cause label, and their values ​​in the fault backtracing analysis report.

[0290] Specifically, it also includes a visualization display and decision support module, which is used after the fault backtracking analysis report is generated.

[0291] Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols;

[0292] Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window;

[0293] At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence;

[0294] When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

[0295] The following takes the B-phase fault of a 500kV XX line at X:X:X on X / X / 202X as an example to illustrate the specific application process of the transmission line fault retrospective analysis system of the present invention.

[0296] The data acquisition and fusion platform first collects raw data from each monitoring device:

[0297] The lightning location system recorded seven lightning events within ±2 minutes before and after the fault, including one with a maximum current of –102.8 kA near tower 218 at 21:26:08.791 s.

[0298] The distributed traveling wave monitoring equipment uploaded the arrival time of the traveling wave that triggered the trip and located the fault point between towers 187 and 512, and finally located it near tower 218;

[0299] Meteorological monitoring equipment obtains thunderstorm weather data for the surrounding area during this period: temperature 20°C–32°C, southeast wind force 1, and relative humidity 85%;

[0300] The video surveillance equipment captured a frame of ground flash footage near tower No. 218.

[0301] The data collection and fusion platform adds unified timestamps and geographic coordinate labels to the above multi-source data, and automatically aggregates them into a single "event cluster" with ±0.5s and 5km as time windows and space windows.

[0302] The multi-source data analysis module performs hierarchical processing on the event cluster:

[0303] The traveling wave layer is automatically identified as a "single-peak waveform," and the main peak time (21:26:08.79s) and the waveform rise time and pulse width are extracted;

[0304] Lightning layer verification: The lightning strike time falls within the traveling wave fault time window, the spatial distance is ≤5km, the intensity is ≥20kA, and the intensity is extracted to be –102.8kA and the precise longitude and latitude;

[0305] The meteorological layer verifies that the thunderstorm conditions meet the wind speed <25m / s (in this case, force 1 wind) but the rainfall is extremely large, and the rainfall accumulation value is extracted;

[0306] The video layer identifies ground lightning scenes in key frames and extracts the frame number and the abnormal label "ground lightning".

[0307] The fault judgment module performs multi-condition fusion judgment based on the extracted multi-source characterization features:

[0308] The traveling wave is single-peaked, the lightning event exists and is strong, which is a "lightning-type fault";

[0309] The "ground lightning" scene was captured in the video to further confirm the cause of the fault;

[0310] Final judgment: The fault type was lightning strike, the fault location was near tower 218, and the fault time was 21:26:08.79s.

[0311] The fault backtracking module automatically generates a backtracking analysis report based on the fault type and location results. The report includes:

[0312] Fault summary: 2024-08-17 21:26:08.79s XX line B phase tripped and reclosed successfully;

[0313] Judgment result: Lightning strike type fault, fault location: Tower 218, latitude and longitude ××.×°N / ××.×°E;

[0314] Supporting data: Single peak characteristics of traveling waves, lightning intensity –102.8kA, video frame of “ground lightning”;

[0315] Fault cause label: "Lightning strike", along with the characteristic features and values ​​that triggered the label.

[0316] The report is exported in PDF format and sent to the operation and maintenance mailbox.

[0317] After the report is generated, the platform starts the visualization display and decision support module:

[0318] Map view: The location of tower 218 is marked with a lightning icon on the line topology and highlighted in red;

[0319] Timeline: The fault point is marked at 21:26, and the user can drag the slider to view the events before and after ±1 minute;

[0320] Multi-source details: Click on the map or timeline to draw the traveling wave waveform, lightning bubble diagram (–102.8kA), rainfall time series curve, and key video frame "ground lightning" image at that moment;

[0321] Interactive linkage: When users slide on the timeline or map, all views are refreshed synchronously, intuitively presenting the fault evolution process.

[0322] Through this embodiment, the system realizes a complete closed loop from multi-source data aggregation, layered verification, fusion judgment to report generation and visual display, providing operation and maintenance personnel with a one-stop, efficient and traceable fault backtracking analysis solution.

[0323] Example 3

[0324] The present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the transmission line fault backtracking analysis method described in the above technical solution.

[0325] Example 4

[0326] The present invention also provides an electronic device, comprising: a memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the transmission line fault backtracking analysis method described in the above technical solution by executing the computer instructions.

[0327] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0328] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0329] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0330] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 The steps for the function specified in one or more boxes.

[0331] The contents not described in detail in this specification belong to the prior art known to those skilled in the art.

Claims

1. A transmission line fault retrospective analysis method, characterized by: The following steps are involved: Perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features; The multi-source fault characterization features are logically combined according to the preset judgment conditions of each fault type to obtain the fault type and determine the fault location and time.

2. The transmission line fault retrospective analysis method according to claim 1, characterized in that: Before layering, it also includes: Lightning events, traveling wave waveforms, meteorological parameters and video image data are annotated with unified timestamps and geographic locations, and the annotated multi-source data are aggregated into several event clusters based on the preset time window and spatial range.

3. The transmission line fault retrospective analysis method according to claim 2, characterized in that: The layered processing includes: Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data; For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information; Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics; Identify key frame anomalies in video image data and extract fault scene types.

4. The transmission line fault retrospective analysis method according to claim 3, characterized in that: The multi-condition fusion determination process includes: When all of the following conditions are met, it is determined to be a lightning-type fault: The traveling wave waveform type is single peak or multi-peak; A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

5. The transmission line fault retrospective analysis method according to claim 3, characterized in that: The multi-condition fusion determination process includes: When all of the following conditions are met, it is determined to be a meteorological fault: The traveling wave waveform type is an oscillating waveform or an irregular waveform; The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window; The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold; Rain or strong winds were detected in the video keyframes.

6. The transmission line fault retrospective analysis method according to claim 3, characterized in that: The multi-condition fusion determination process includes: When all of the following conditions are met, it is determined to be an external force fault: The traveling wave waveform type is a single peak waveform or an irregular waveform; The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold; Meteorological parameters do not meet the preset wind speed and rainfall thresholds; External force damage was detected in the video key frames.

7. The transmission line fault retrospective analysis method according to any one of claims 4 to 6, characterized in that: The multi-condition fusion determination process further includes: When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

8. The transmission line fault retrospective analysis method according to claim 1, characterized in that: Also includes: Generate a fault backtracking analysis report based on the fault type, fault location and time.

9. The transmission line fault retrospective analysis method according to claim 1, characterized in that: The following steps are also included: According to the fault type determined by the multi-condition fusion judgment step, the corresponding fault cause label is automatically generated based on the mapping relationship between the multi-source fault characterization features used and the preset cause; The fault cause label, the characteristic features that trigger the fault cause label, and their values ​​are included in the fault backtracking analysis report.

10. The transmission line fault retrospective analysis method according to claim 1, characterized in that: After the fault backtracking analysis report is generated, the following visual display steps are also included: Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols; Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window; At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence; When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

11. A transmission line fault retrospective analysis system, characterized by: include: Multi-source data analysis module, used to perform layered processing on lightning events, traveling wave waveforms, meteorological parameters, and video image data of transmission lines to obtain corresponding fault characterization features; The fault diagnosis module is used to logically combine the multi-source fault characterization features according to the preset judgment conditions of each fault type, obtain the fault type and determine the fault location and time.

12. The power transmission line fault retrospective analysis system according to claim 11, characterized in that: It also includes a data acquisition and fusion platform, which is used to mark lightning events, traveling wave waveforms, meteorological parameters and video image data with unified timestamps and geographic locations, and aggregate the marked multi-source data into several event clusters based on preset time windows and spatial ranges and send them to the multi-source data analysis module.

13. The power transmission line fault retrospective analysis system according to claim 11, characterized in that: The hierarchical processing of the multi-source data analysis module includes: Automatically identify the preset type of traveling wave waveform data and extract the traveling wave fault time window and corresponding waveform data; For lightning event data, verify whether the lightning strike occurrence time matches the traveling wave fault time window, whether the spatial distance between the lightning strike point and the transmission line is within a set range, and whether the lightning strike intensity reaches a preset threshold; and extract the verified lightning strike intensity value and spatial location and multiple grounding point lightning strike information; Validate thresholds on meteorological parameter data and extract wind speed and / or rainfall metrics; Identify key frame anomalies in video image data and extract fault scene types.

14. The power transmission line fault retrospective analysis system according to claim 13, characterized in that: The multi-condition fusion judgment process of the multi-source data analysis module includes: When all of the following conditions are met, it is determined to be a lightning-type fault: The traveling wave waveform type is single peak or multi-peak; A lightning event exists, the lightning strike time matches the traveling wave fault time window, the spatial distance between the lightning strike point and the transmission line does not exceed the set threshold, and the lightning strike intensity reaches the preset threshold.

15. The transmission line fault retrospective analysis system according to claim 13, characterized in that: The multi-condition fusion judgment process of the multi-source data analysis module includes: When all of the following conditions are met, it is determined to be a meteorological fault: The traveling wave waveform type is an oscillating waveform or an irregular waveform; The lightning data is empty, or the lightning strike intensity is lower than the preset threshold and the lightning strike occurrence time does not match the traveling wave fault time window; The wind speed or rainfall in the meteorological parameters meets the corresponding set threshold; Rain or strong winds were detected in the video keyframes.

16. The power transmission line fault tracing analysis system according to claim 13, characterized in that: The multi-condition fusion judgment process of the multi-source data analysis module includes: When all of the following conditions are met, it is determined to be an external force fault: The traveling wave waveform type is a single peak waveform or an irregular waveform; The lightning event is empty or does not meet the lightning strike time and the traveling wave fault time window, or the spatial distance between the lightning strike point and the transmission line exceeds the set threshold; Meteorological parameters do not meet the preset wind speed and rainfall thresholds; External force damage was detected in the video key frames.

17. The power transmission line fault retrospective analysis system according to any one of claims 14 to 16, characterized in that: The multi-condition fusion process steps of the multi-source data analysis module also include: When the characterization features of multi-source faults simultaneously meet the judgment conditions of two or more fault types, the preset machine learning model or expert system is called for further analysis and determined to be a complex fault.

18. The power transmission line fault retrospective analysis system according to claim 10, characterized in that: Also includes: The fault backtracing module is used to generate a fault backtracing analysis report based on the fault type, fault location and time.

19. The power transmission line fault retrospective analysis system according to claim 18, characterized in that: The fault backtracking module is also used to automatically generate corresponding fault cause labels based on the fault type determined by multi-condition fusion and the mapping relationship between the multi-source fault characterization features used and the preset causes; The fault cause label, the characteristic features that trigger the fault cause label, and their values ​​are included in the fault backtracking analysis report.

20. The power transmission line fault retrospective analysis system according to claim 18, characterized in that: It also includes a visualization and decision support module for generating the fault backtracking analysis report. Mark the location of each fault event on a geographical map and distinguish the fault type with different symbols; Presents the occurrence time distribution of fault events on the time axis and allows users to select any time window; At the location or time window of the selected fault event, the corresponding traveling wave waveform curve, lightning strike intensity and spatial distribution diagram, wind speed / rainfall time series curve, and the corresponding video key frame are drawn in sequence; When the user selects or slides in the map or timeline, all the above views are refreshed synchronously to show the spatiotemporal evolution of the fault and the results of multi-source analysis.

21. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for tracing back fault analysis of a power transmission line according to any one of claims 1 to 10 is implemented.

22. An electronic device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the transmission line fault tracing analysis method according to any one of claims 1 to 10 by executing the computer instructions.

Images