Holographic diagnosis fault positioning system for low-voltage distribution line

Abstract

The invention discloses a holographic diagnosis fault positioning system for a low-voltage distribution line, and relates to the field of holographic diagnosis, and the system comprises the steps: obtaining the electrical, environment and physical states of the line and equipment operation parameters in real time through a data collection module; after noise is eliminated and data is complemented, a calibration module carries out precision correction based on a standard model library; the holographic data fusion module adopts a three-level progressive strategy to integrate multi-source standardized data; the fault feature extraction module identifies fault features from the fused data in combination with deep learning; the holographic diagnosis analysis module fuses a mechanism analysis model, a machine learning model and a rule reasoning model to realize multi-dimensional diagnosis, and the precise positioning module performs fault point calculation by using spatial-temporal characteristics; and the reliability of the diagnosis result is ensured through a four-stage closed loop verification mechanism. The method has the advantages that based on full-link multi-dimensional holographic data support, multi-model collaborative diagnosis and multi-algorithm fusion positioning are matched with four-stage closed-loop verification, and comprehensive diagnosis and reliable positioning of low-voltage distribution line faults are achieved.

Description

Technical Field

[0001] This invention relates to the field of holographic diagnostics, and in particular to a holographic diagnostic fault location system for low-voltage power distribution lines. Background Technology

[0002] With the rapid growth of new loads such as distributed energy and electric vehicles, traditional distribution networks face problems such as high failure rates, low fault location efficiency, and expanded impact of power outages. Statistics show that distribution network faults account for more than 80% of power system accidents, and traditional methods such as manual inspection and segmented power restoration require an average of 2-4 hours to locate the fault point, seriously affecting the reliability of power supply.

[0003] The core weaknesses of current low-voltage distribution line fault diagnosis and location systems lie in four main dimensions: data perception, fusion analysis, location accuracy, and verification mechanisms. Data acquisition is often limited to single electrical parameters, lacking multi-dimensional data such as environmental and line physical conditions. Furthermore, fusion technology is rudimentary, making it difficult to achieve cross-dimensional spatiotemporal alignment and deep correlation, and thus unable to construct a comprehensive holographic view of line operation. Diagnostic models often rely on single algorithms or traditional signal processing methods, lacking multi-model collaborative mechanisms, resulting in low recognition rates for complex faults and high-resistance faults, and susceptibility to interference from complex scenarios such as distributed power source integration. Location algorithms have poor adaptability, exhibiting significant errors in complex topologies such as branch lines and mixed lines, making accurate location difficult. Simultaneously, a comprehensive end-to-end verification mechanism is generally lacking, compromising the reliability of diagnostic and location results, and relying heavily on manual assessment, leading to low fault handling efficiency and failing to meet the precise operation and maintenance needs of complex distribution networks. Summary of the Invention

[0004] To improve the existing system, a holographic diagnostic fault location system for low-voltage power distribution lines is provided. This method relies on multi-dimensional holographic data across the entire power supply chain, and achieves comprehensive and accurate diagnosis and reliable location of faults in low-voltage power distribution lines through multi-model collaborative diagnosis and multi-algorithm fusion positioning, coupled with four-level closed-loop verification.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A holographic fault location system for low-voltage power distribution lines includes: Data acquisition module: used to collect real-time, multi-dimensional operational data of low-voltage power distribution lines, construct holographic data for fault diagnosis, and the collection scope covers line electrical parameters, environmental parameters, line physical state parameters and equipment operating state parameters; Data preprocessing module: Connected to the data acquisition module, it is used to remove noise, filter redundancy, and complete the integrity of the acquired raw data; Data calibration module: Built-in multi-dimensional standard calibration model library, preset corresponding calibration parameters and correction algorithms for different types of collected data, and perform accuracy calibration and correction of deviations on pre-processed clean data; Holographic data fusion module: It adopts a three-level progressive holographic fusion strategy to perform cross-dimensional and full-element holographic fusion processing on the calibrated standardized data to construct a holographic dataset that comprehensively depicts the line's operating status; Fault Feature Extraction Module: Constructs a multi-type fault feature template library, and uses a hybrid feature extraction method combining deep learning and traditional signal processing to extract feature data related to various faults from holographic fusion data; Holographic Diagnostic Analysis Module: Integrates an analytical model based on fault mechanism, a machine learning diagnostic model based on deep learning, and a rule-based reasoning model based on expert experience. Through the synergistic linkage of these three models, it performs full-dimensional and multi-perspective holographic diagnosis of low-voltage power distribution line faults. Precise positioning calculation module: It adopts a multi-algorithm fusion positioning strategy, combines the temporal and spatial features of holographic data to complete the positioning calculation, and realizes the precise positioning of the fault point; Diagnostic result verification module: A four-level closed-loop verification mechanism is used to verify the diagnostic conclusions and fault location results in a multi-dimensional and full-link manner.

[0006] Preferably, the data acquisition module specifically includes: Multi-type sensing units: collect line electrical parameters, environmental parameters, line physical state parameters and equipment operating status parameters, specifically including: three-phase voltage, three-phase current, zero-sequence current, line active power, reactive power, power factor electrical parameters; temperature, humidity, wind speed, precipitation, ice thickness environmental parameters; line joint temperature, insulator pollution level, conductor sag line physical state parameters; and operating status parameters of distribution transformers, circuit breakers, fuses equipment. Hybrid communication unit: LoRa self-organizing network transmission is used for short distances, and Ethernet is used to access the distribution network backbone for long distances; Clock synchronization unit: integrates Beidou time synchronization module, dynamically calibrates the timestamps of each acquisition terminal through NTP protocol.

[0007] Preferably, the data preprocessing module specifically includes: Noise Removal Unit: Identifies abnormal noise data through a multi-threshold screening mechanism, adaptively adjusts the filtering window according to the intensity of data fluctuations, and removes noise caused by electromagnetic interference from power distribution lines; Data normalization unit: By distinguishing the dimensional attributes of different types of data such as electrical, environmental and equipment status data, a hierarchical scaling strategy is used to map the data to the [0,1] interval and record the dimensional conversion mapping relationship; Time-series interpolation completion unit: Based on the historical trend of data changes and similar working conditions data of the same period, a piecewise interpolation strategy is used to complete the missing data; Data quality verification unit: Constructs a three-dimensional verification index system of integrity, consistency and accuracy, performs full verification on preprocessed data, and triggers a backtracking and reprocessing process for unqualified data.

[0008] Preferably, the data calibration module specifically includes: Standard calibration model library unit: It has three major categories of calibration models: electrical, environmental and equipment status. It presets calibration reference parameters for different data acquisition terminal models and line operating conditions, and stores historical calibration data and deviation correction data. Scene matching and model selection unit: Extracts the acquisition scene information and acquisition terminal identification information of the preprocessed data, and matches it with the corresponding standard calibration model; Deviation correction unit: Calculates the deviation between the measured data and the standard model reference value in real time, and dynamically adjusts the correction coefficient based on the running time of the acquisition terminal and changes in ambient temperature and humidity to complete deviation compensation; Calibration result verification unit: Constructs dual verification indicators of accuracy and consistency, performs full verification of calibrated data, triggers backtracking and recalibration process for unqualified data, and attaches calibration accuracy rating label to qualified data.

[0009] Preferably, the holographic data fusion module specifically includes: Data layer fusion unit: performs time axis alignment and redundancy removal on standardized data from different acquisition nodes of the same type, and generates a continuous data sequence covering the entire link through a sliding window aggregation algorithm; Feature layer fusion unit: Extracts time-domain, frequency-domain and time-frequency-domain features from data of each dimension, establishes cross-dimensional feature associations through feature correlation analysis, and generates a unified feature set; Decision-making fusion unit: It assigns weights and resolves conflicts based on the preliminary state judgment results of data from each dimension, and outputs a comprehensive state assessment result after integrating complementary information. Fusion process control unit: coordinates the collaborative work of the three-level fusion units, records fusion process parameters, and triggers a backtracking and re-fusion mechanism for abnormal fusion data.

[0010] Preferably, the fault feature extraction module specifically includes: Fault Feature Template Library Unit: Built-in short circuit, grounding, and overload fault feature subsets, storing typical time-domain sudden change thresholds, frequency-domain distortion intervals, and feature change time sequence patterns for each fault, and associating fault causes with feature mapping relationships; Multi-domain feature mining unit: integrates time domain, frequency domain, and deep abstract feature extraction sub-modules. The time domain sub-module captures the peak value and duration of data mutations, the frequency domain sub-module identifies fault frequency distortion features, and the deep sub-module mines non-explicit correlation features. Feature matching and filtering unit: Employs a feature similarity comparison algorithm to accurately match multi-domain mined features with a template library, generating targeted fault feature vectors; Feature quality control unit: performs integrity and consistency checks on the generated feature vectors, and triggers a backtracking and re-mining process for unqualified features.

[0011] Preferably, the holographic diagnostic analysis module specifically includes: Multi-model fusion diagnostic engine unit: integrates three sub-modules: analytical model, machine learning diagnostic model and rule reasoning model. The analytical model constructs a logical judgment link based on the fault mechanism to lock the range of fault types. The machine learning model is equipped with a deep training network to mine hidden correlation information in holographic data. The rule reasoning model integrates the domain expert experience base to form a complementary diagnostic link. Model Coordination Scheduling Unit: Employs a dynamic weight allocation mechanism to adjust the output weights of each model based on the complexity of the fault scenario, integrates differing conclusions through a conflict resolution algorithm, and outputs a unified diagnostic result; Learning and updating unit: Real-time absorption of new fault data and diagnostic experience, and dynamic optimization of model parameters and weight configuration.

[0012] Preferably, the precise positioning calculation module specifically includes: Preliminary location calculation unit: Based on the electrical parameters and inherent parameters of the line in the holographic fusion data, a fault impedance analysis link is constructed to obtain the initial distance between the fault point and the acquisition node; Positioning result correction unit: Based on the propagation time and reflection characteristics of the fault traveling wave, calculate the propagation speed and time difference of the traveling wave, perform deviation compensation on the preliminary positioning result, and obtain the corrected accurate distance data; Spatial topology mapping unit: Converts the corrected fault distance into geographic coordinates, associates them with the corresponding line section and equipment identification information, generates distance-coordinate-equipment three-dimensional association data, and accurately maps the fault point space; Complex Scene Adaptation Unit: For complex scenes such as line branches and intersections, dynamically match the corresponding scene positioning algorithm parameters.

[0013] Preferably, the diagnostic result verification module specifically includes: Data consistency verification unit: Extract the change curves of holographic fusion data before and after the fault, compare the matching degree between the diagnostic conclusions and the data mutation characteristics and trend anomalies, screen unqualified diagnostic results and mark the deviation type; Model cross-validation unit: By reanalyzing the same fault data through the parsing, machine learning, and rule-based reasoning models of the holographic diagnostic core, the consistency coefficient of the diagnostic conclusions of each model is calculated to verify the stability of the diagnosis.

[0014] Spatial correlation verification unit: Combining real-time monitoring images of the line, UAV inspection data and equipment operation status feedback information, the fault point location results are verified in the field. Historical Case Verification Unit: The current fault's feature vector, diagnostic conclusion, and location information are compared with the historical fault case database to analyze the correlation between fault type, cause, location, and influencing factors, and to verify the rationality of the diagnostic and location results.

[0015] Compared with the prior art, the advantages of the present invention are: A comprehensive data system and multi-dimensional collaborative diagnostic and localization mechanism are constructed across the entire power distribution line. Through multi-dimensional data acquisition and three-level progressive fusion, a complete characterization of the line's operating status is achieved. Hybrid feature extraction combining deep learning and traditional signal processing, coupled with the collaborative operation of three diagnostic models, ensures comprehensive and accurate fault diagnosis. Precise fault location is achieved through multi-algorithm fusion localization strategies and time-series-spatial feature mining. A four-level closed-loop verification mechanism comprehensively guarantees the reliability of diagnostic and localization results. Simultaneously, the system integrates efficient data preprocessing, dynamic calibration, and flexible communication and time-series synchronization capabilities, further improving data quality and transmission stability, providing a comprehensive, high-precision, and highly reliable solution for fault diagnosis and localization of low-voltage power distribution lines. Attached Figure Description

[0016] Figure 1 This is a schematic diagram of the system proposed in this invention; Figure 2 This is a diagram of the data acquisition module proposed in this invention; Figure 3 This is a diagram of the data preprocessing module proposed in this invention; Figure 4 This is a diagram of the data calibration module proposed in this invention; Figure 5 This is a diagram of the holographic data fusion module proposed in this invention; Figure 6 This is a diagram of the fault feature extraction module proposed in this invention; Figure 7 This is a diagram of the holographic diagnostic analysis module proposed in this invention; Figure 8 This is a diagram of the precise positioning calculation module proposed in this invention; Figure 9 This is a diagram of the diagnostic result verification module proposed in this invention. Detailed Implementation

[0017] The following description is intended to disclose the invention and enable those skilled in the art to implement it. The preferred embodiments described below are merely examples, and other obvious variations will occur to those skilled in the art.

[0018] See Figure 1 As shown, a holographic fault location system for low-voltage power distribution lines includes: Data acquisition module: used to collect real-time, multi-dimensional operational data of low-voltage power distribution lines, construct holographic data for fault diagnosis, and the collection scope covers line electrical parameters, environmental parameters, line physical state parameters and equipment operating state parameters; Data preprocessing module: Connected to the data acquisition module, it is used to remove noise, filter redundancy, and complete the integrity of the acquired raw data; Data calibration module: Built-in multi-dimensional standard calibration model library, preset corresponding calibration parameters and correction algorithms for different types of collected data, and perform accuracy calibration and correction of deviations on pre-processed clean data; Holographic data fusion module: It adopts a three-level progressive holographic fusion strategy to perform cross-dimensional and full-element holographic fusion processing on the calibrated standardized data to construct a holographic dataset that comprehensively depicts the line's operating status; Fault Feature Extraction Module: Constructs a multi-type fault feature template library, and uses a hybrid feature extraction method combining deep learning and traditional signal processing to extract feature data related to various faults from holographic fusion data; Holographic Diagnostic Analysis Module: Integrates an analytical model based on fault mechanism, a machine learning diagnostic model based on deep learning, and a rule-based reasoning model based on expert experience. Through the synergistic linkage of these three models, it performs full-dimensional and multi-perspective holographic diagnosis of low-voltage power distribution line faults. Precise positioning calculation module: It adopts a multi-algorithm fusion positioning strategy, combines the temporal and spatial features of holographic data to complete the positioning calculation, and realizes the precise positioning of the fault point; Diagnostic result verification module: A four-level closed-loop verification mechanism is used to verify the diagnostic conclusions and fault location results in a multi-dimensional and full-link manner.

[0019] See Figure 2 As shown, the data acquisition module specifically includes: Multi-type sensing units: collect line electrical parameters, environmental parameters, line physical state parameters and equipment operating status parameters, specifically including: three-phase voltage, three-phase current, zero-sequence current, line active power, reactive power, power factor electrical parameters; temperature, humidity, wind speed, precipitation, ice thickness environmental parameters; line joint temperature, insulator pollution level, conductor sag line physical state parameters; and operating status parameters of distribution transformers, circuit breakers, fuses equipment. Hybrid communication unit: LoRa self-organizing network transmission is used for short distances, and Ethernet is used to access the distribution network backbone for long distances; Clock synchronization unit: integrates Beidou time synchronization module, dynamically calibrates the timestamps of each acquisition terminal through NTP protocol.

[0020] Specifically, based on the topology, load distribution, and high-fault areas of low-voltage power distribution lines, a distributed deployment strategy of "full coverage of the main line + precise supplementary points on branch lines" is adopted. For key equipment such as distribution transformers and circuit breakers, dedicated acquisition units are installed at both the outgoing and incoming ends of the equipment to ensure that the acquisition range covers the entire line link without any data acquisition blind spots. The connection between the acquisition terminal and the line and equipment adopts an insulated piercing joint to avoid damaging the line insulation layer and ensure the safety of the deployment process. The terminal starts data collection according to a preset collection cycle; electrical parameters are collected by sensing line current and voltage signals through a built-in high-precision Hall sensor, which are then converted into standard electrical signals by a signal conditioning circuit; environmental parameters are collected in real time by integrating dedicated sensors such as temperature and humidity sensors, wind speed sensors, and ultrasonic icing thickness detectors to collect data on the surrounding environment of the line; line physical status parameters are collected by using an infrared temperature measurement module to collect joint temperature, using a high-definition camera with an image recognition module to capture the degree of insulator contamination, and using a laser rangefinder to detect conductor sag; equipment status parameters are obtained by reading data from the built-in monitoring chips of distribution transformers and circuit breakers to obtain information such as equipment operating current, temperature, and insulation status.

[0021] See Figure 3 As shown, the data preprocessing module specifically includes: Noise Removal Unit: Identifies abnormal noise data through a multi-threshold screening mechanism, adaptively adjusts the filtering window according to the intensity of data fluctuations, and removes noise caused by electromagnetic interference from power distribution lines; Data normalization unit: By distinguishing the dimensional attributes of different types of data such as electrical, environmental and equipment status data, a hierarchical scaling strategy is used to map the data to the [0,1] interval and record the dimensional conversion mapping relationship; Time-series interpolation completion unit: Based on the historical trend of data changes and similar working conditions data of the same period, a piecewise interpolation strategy is used to complete the missing data; Data quality verification unit: Constructs a three-dimensional verification index system of integrity, consistency and accuracy, performs full verification on preprocessed data, and triggers a backtracking and reprocessing process for unqualified data.

[0022] Specifically, various noise interferences in the raw data, such as electromagnetic interference, sensor errors, and transmission link noise, are addressed using a classification-based adaptive noise reduction strategy. For high-frequency electrical parameter data, an adaptive sliding window filter is used to dynamically adjust the window size, identifying and removing impulse noise and random noise. For low-frequency environmental parameters and equipment status data, trend fitting noise reduction is employed, using a time-series fitting curve to remove outliers that deviate from the threshold. All noise removal operations are logged in detail, including the noise time, location, and the basis for removal. In multi-dimensional data normalization processing, different methods are used according to the differences in data types; by establishing a data type-dimensional mapping table, electrical parameters are linearly normalized and compressed to the [0,1] interval, while retaining relative relationships; environmental parameters and equipment status parameters are standardized and normalized to eliminate distribution differences; and dimensionless parameters are directly verified to be within a reasonable range. For missing time-series data, a tiered completion strategy is implemented: for short-term missing data, linear interpolation of adjacent data is used; for long-term missing data, a time-series matching model is constructed by combining historical data and similar operating conditions, and completion is achieved through trend migration; continuous missing key nodes will trigger alarms and mark the reliability level; after completion, consistency verification is used to ensure that the data trend is continuous and reasonable, and finally a complete and reliable data sequence is formed.

[0023] See Figure 4 As shown, the data calibration module specifically includes: Standard calibration model library unit: It has three major categories of calibration models: electrical, environmental and equipment status. It presets calibration reference parameters for different data acquisition terminal models and line operating conditions, and stores historical calibration data and deviation correction data. Scene matching and model selection unit: Extracts the acquisition scene information and acquisition terminal identification information of the preprocessed data, and matches it with the corresponding standard calibration model; Deviation correction unit: Calculates the deviation between the measured data and the standard model reference value in real time, and dynamically adjusts the correction coefficient based on the running time of the acquisition terminal and changes in ambient temperature and humidity to complete deviation compensation; Calibration result verification unit: Constructs dual verification indicators of accuracy and consistency, performs full verification of calibrated data, triggers backtracking and recalibration process for unqualified data, and attaches calibration accuracy rating label to qualified data.

[0024] Specifically, by combining the matched calibration model, real-time self-calibration parameters of the acquisition terminal are extracted, and data system deviations are calculated. For electrical parameters, the amplitude and phase deviations of parameters such as voltage and current are determined by comparing the acquired data with the reference values ​​of the terminal's built-in standard resistor and standard voltage source. For environmental parameters, the sensitivity deviation of environmental sensors is calculated based on the difference between the data from the same period's standard meteorological station and the acquired data. For equipment status parameters, the drift deviation of the equipment's operating status parameters is calculated by combining the equipment's factory standard parameters with the current acquired data. According to the deviation type, the corresponding correction strategy is called to perform point-by-point deviation compensation on the data, completing the preliminary calibration. After the initial calibration is completed, a two-stage accuracy verification is initiated. The first stage is the benchmark comparison verification, which compares the calibrated data with the benchmark data output by the standard calibration source. If the deviation exceeds the preset threshold, the calibration parameters are readjusted and the correction is repeated. The second stage is the temporal stability verification, which analyzes the fluctuation range of the calibrated data within 10 consecutive acquisition cycles to ensure that the data stability meets the requirements. The module records the running time of the acquisition terminal and environmental change data in real time, and dynamically updates the aging correction coefficient and environmental interference coefficient in the calibration model to form a personalized dynamic calibration parameter library. If the verification is successful, high-precision standardized data is output.

[0025] See Figure 5 As shown, the holographic data fusion module specifically includes: Data layer fusion unit: performs time axis alignment and redundancy removal on standardized data from different acquisition nodes of the same type, and generates a continuous data sequence covering the entire link through a sliding window aggregation algorithm; Feature layer fusion unit: Extracts time-domain, frequency-domain and time-frequency-domain features from data of each dimension, establishes cross-dimensional feature associations through feature correlation analysis, and generates a unified feature set; Decision-making fusion unit: It assigns weights and resolves conflicts based on the preliminary state judgment results of data from each dimension, and outputs a comprehensive state assessment result after integrating complementary information. Fusion process control unit: coordinates the collaborative work of the three-level fusion units, records fusion process parameters, and triggers a backtracking and re-fusion mechanism for abnormal fusion data.

[0026] Specifically, the data layer holographic aggregation and integration adopts a "type-based - full-link" aggregation strategy to process standardized data; it classifies data by data type and performs sliding window aggregation on time-series data from different acquisition nodes under the same type to eliminate duplicate and redundant data; for the acquisition data of the trunk line and branch line, it establishes a spatial mapping relationship through the line topology association algorithm, fills in the data breakpoints of the link, and forms a continuous data sequence covering the entire length and all nodes of the line; it adds spatial identifiers to the aggregated data to achieve accurate "time-space" dual-dimensional association of data, and completes the holographic integration of the data layer with full spatial and temporal coverage; The feature-layer holographic correlation mining is based on the data-layer aggregation results, initiating multi-dimensional feature extraction and correlation mapping; time-domain features, frequency-domain features, and time-frequency domain joint features are extracted from various types of data to form a multi-dimensional feature set; through feature correlation analysis algorithms, the intrinsic correlation between features of different dimensions is mined, such as the correlation between icing thickness and line impedance, and the correlation between ambient temperature and equipment temperature, to construct a feature correlation map; closely correlated features are fused and dimensionality reduced, redundant features are eliminated, and core complementary features are retained to form a holographic feature set that can comprehensively reflect the line's operating status; The decision-making layer uses an improved evidence theory fusion algorithm to perform decision fusion on the preliminary state judgment results of data from various dimensions. Based on the feature layer holographic feature set, the preliminary judgment and confidence level of the line status corresponding to each dimension of data are obtained through the basic discriminant model. Then, each preliminary judgment result is used as evidence input, and the complementary information of evidence from different dimensions is integrated through evidence synthesis rules to solve the uncertainty problem of single-dimensional judgment. Finally, the fusion result is optimized and corrected by combining the line's historical operation data and typical operating condition characteristics, and the output is a holographic fusion dataset that can truly and comprehensively depict the line's operating status, while simultaneously marking the data reliability level.

[0027] See Figure 6 As shown, the fault feature extraction module specifically includes: Fault Feature Template Library Unit: Built-in short circuit, grounding, and overload fault feature subsets, storing typical time-domain sudden change thresholds, frequency-domain distortion intervals, and feature change time sequence patterns for each fault, and associating fault causes with feature mapping relationships; Multi-domain feature mining unit: integrates time domain, frequency domain, and deep abstract feature extraction sub-modules. The time domain sub-module captures the peak value and duration of data mutations, the frequency domain sub-module identifies fault frequency distortion features, and the deep sub-module mines non-explicit correlation features. Feature matching and filtering unit: Employs a feature similarity comparison algorithm to accurately match multi-domain mined features with a template library, generating targeted fault feature vectors; Feature quality control unit: performs integrity and consistency checks on the generated feature vectors, and triggers a backtracking and re-mining process for unqualified features.

[0028] Specifically, the built-in multi-type fault feature template library is retrieved. This template library is constructed by integrating massive historical fault data based on the occurrence mechanism of common faults such as short circuit, grounding, overload, poor contact, and insulation aging in low-voltage power distribution lines. It includes typical time-domain features, frequency-domain features, and time-frequency domain joint features corresponding to various faults. According to the line type and equipment configuration information of the target dataset, the corresponding fault feature template subset is matched, and the feature extraction parameters and filtering thresholds are initialized. A hybrid strategy combining deep learning and traditional signal processing is adopted to carry out full-dimensional feature mining. Basic features are extracted through traditional signal processing methods: wavelet transform is used to decompose the target data at multiple scales to capture the signal mutation features at the moment of fault occurrence and locate the precise time point of the mutation; Fourier transform is used to convert the time domain data to the frequency domain to extract the frequency distortion features under fault conditions and identify abnormal harmonic components and main frequency shifts; then, a convolutional neural network model is launched to perform deep feature mining on the preprocessed time series data, extracting the non-explicit features hidden in the data and forming a multi-dimensional initial feature set covering basic and deep features.

[0029] See Figure 7 As shown, the holographic diagnostic analysis module specifically includes: Multi-model fusion diagnostic engine unit: integrates three sub-modules: analytical model, machine learning diagnostic model and rule reasoning model. The analytical model constructs a logical judgment link based on the fault mechanism to lock the range of fault types. The machine learning model is equipped with a deep training network to mine hidden correlation information in holographic data. The rule reasoning model integrates the domain expert experience base to form a complementary diagnostic link. Model Coordination Scheduling Unit: Employs a dynamic weight allocation mechanism to adjust the output weights of each model based on the complexity of the fault scenario, integrates differing conclusions through a conflict resolution algorithm, and outputs a unified diagnostic result; Learning and updating unit: Real-time absorption of new fault data and diagnostic experience, and dynamic optimization of model parameters and weight configuration.

[0030] Specifically, an analytical model based on fault mechanism is initiated, and the occurrence mechanism and evolution law of different faults in low-voltage distribution lines are combined to conduct targeted analysis of specific fault feature vectors. By matching the inherent correspondence between fault features and fault mechanisms, such as the current mutation characteristics corresponding to short-circuit faults and the zero-sequence current anomaly characteristics corresponding to ground faults, the approximate type range and possible causes of faults are quickly identified, and preliminary diagnostic results and confidence scores are output. The system invokes a deep learning-based machine learning diagnostic model, using a preprocessed subset of holographic fusion data and the preliminary diagnostic results output by the analytical model as joint inputs. Based on the feature-fault mapping relationship formed by training on massive historical fault data, the model deeply mines the hidden non-explicit correlation information in the holographic data, such as the correlation between environmental temperature and humidity and insulation aging faults, and the evolution correlation between line load changes and overload faults. The system further refines and identifies the fault types identified in the preliminary diagnosis, accurately determines the specific fault type and core inducing factors, and outputs refined diagnostic results and multi-dimensional confidence assessment indicators. The system initiates a rule-based reasoning model based on expert experience, drawing upon the built-in domain expert rule base. It then combines relevant information from the holographic fusion data, such as the line's operational years, equipment maintenance records, and real-time environmental parameters, to supplement and validate the refined diagnostic results output by the machine learning model. For example, if the diagnosis is an insulation aging fault, it needs to verify whether the line's operational years have reached the peak aging period and whether the recent environmental humidity exceeds the standard. Simultaneously, the system evaluates the rationality and completeness of the diagnostic results, corrects the bias of single-model diagnoses, and improves supplementary information such as the fault's impact range and development trend.

[0031] The diagnostic results of the analytical model, machine learning model and rule reasoning model are integrated by adopting a weighted fusion strategy. The weights are dynamically allocated according to the historical accuracy of each model in the diagnosis of different fault types, and the final holographic diagnostic conclusion is generated, which covers all dimensions of information such as the specific type of fault, the core cause, the severity, the scope of impact and the development trend.

[0032] See Figure 8 As shown, the precise positioning calculation module specifically includes: Preliminary location calculation unit: Based on the electrical parameters and inherent parameters of the line in the holographic fusion data, a fault impedance analysis link is constructed to obtain the initial distance between the fault point and the acquisition node; Positioning result correction unit: Based on the propagation time and reflection characteristics of the fault traveling wave, calculate the propagation speed and time difference of the traveling wave, perform deviation compensation on the preliminary positioning result, and obtain the corrected accurate distance data; Spatial topology mapping unit: Converts the corrected fault distance into geographic coordinates, associates them with the corresponding line section and equipment identification information, generates distance-coordinate-equipment three-dimensional association data, and accurately maps the fault point space; Complex Scene Adaptation Unit: For complex scenes such as line branches and intersections, dynamically match the corresponding scene positioning algorithm parameters.

[0033] Specifically, for different fault types, corresponding impedance calculation logic is matched to construct a fault impedance model; using electrical parameter data at the moment of fault occurrence, the equivalent impedance value of the fault circuit is calculated, and combined with the preset impedance parameters per unit length of the line, the distance between the fault point and the nearest data acquisition node is initially estimated. During the process, the influence of load current on impedance calculation needs to be corrected according to the line load distribution; for three-phase unbalanced lines, an unbalance coefficient is introduced for deviation compensation, and the confidence level of the preliminary location results is marked. If the confidence level is lower than the preset threshold, the traveling wave method location process is initiated first. The preliminary location formula using the impedance method is: in, This represents the initial distance between the fault point and the data acquisition node calculated using the impedance method. The equivalent impedance of the faulty circuit. The impedance per unit length of the line. This is the load correction factor; The fault traveling wave signal is extracted from the holographic fusion data. Interference signals are removed through signal filtering. The initial wavefront and reflected wavefront of the traveling wave are identified, and the time difference between the arrival of the two wavefronts at the acquisition node is recorded. Combined with the propagation speed of the traveling wave in the corresponding line medium, the precise distance between the fault point and the acquisition node is calculated. This distance is compared with the preliminary location result of the impedance method. The two data are fused using a weighted average algorithm to correct the location error caused by the line parameter estimation deviation and load fluctuation of the impedance method, so as to obtain a more accurate fault distance value. The system retrieves the spatial topology database of the line and, based on the fused fault distance value and the accurate geographic coordinates of the nearest collected node, performs coordinate conversion along the line direction to obtain the preliminary geographic coordinates of the fault point. For complex topology scenarios such as line branches and ring networks, the system initiates a topology adaptive adjustment algorithm. By matching the segment identifiers and branch node information of the faulty line, the system identifies the specific line branch where the fault is located and corrects the topology deviation in the coordinate conversion process.

[0034] See Figure 9 As shown, the diagnostic result verification module specifically includes: Data consistency verification unit: Extract the change curves of holographic fusion data before and after the fault, compare the matching degree between the diagnostic conclusions and the data mutation characteristics and trend anomalies, screen unqualified diagnostic results and mark the deviation type; Model cross-validation unit: By reanalyzing the same fault data through the parsing, machine learning, and rule-based reasoning models of the holographic diagnostic core, the consistency coefficient of the diagnostic conclusions of each model is calculated to verify the stability of the diagnosis.

[0035] Spatial correlation verification unit: Combining real-time monitoring images of the line, UAV inspection data and equipment operation status feedback information, the fault point location results are verified in the field. Historical Case Verification Unit: The current fault's feature vector, diagnostic conclusion, and location information are compared with the historical fault case database to analyze the correlation between fault type, cause, location, and influencing factors, and to verify the rationality of the diagnostic and location results.

[0036] Specifically, based on the holographic fusion data in the verification dataset, the changing trends of key parameters before and after the fault occurred are compared; for the fault type corresponding to the diagnostic conclusion, the typical data change feature templates of this type of fault are retrieved, and the sudden change amplitude and duration of electrical parameters such as voltage, current, and power, as well as the coordinated change patterns of environmental parameters and line physical state parameters are compared one by one to verify the matching between the diagnostic conclusion and the data feature changes; if the deviation between the data change trend and the typical features corresponding to the diagnostic conclusion exceeds the preset threshold, it is marked as a verification anomaly, and the deviation data points and specific differences are recorded; Activate the analytical model, machine learning model, and rule-based reasoning model in the core analysis module of holographic diagnosis. Input the fault feature data in the verification dataset into the three models for independent diagnostic analysis. Extract the diagnostic results output by each model, including fault type, cause, and confidence level, and compare the consistency of the three: if the diagnostic results of the three are consistent and the confidence levels all meet the standard, the verification of this dimension is deemed successful; if there are differences in the results, combine the historical accuracy of each model in the diagnosis of similar faults to analyze the causes of the differences, and prioritize the results of the model with higher historical accuracy as a reference.

[0037] It should be noted that the order of the above embodiments of the present invention is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some embodiments, multitasking and parallel processing are possible or may be advantageous.

[0038] The various embodiments in this specification are described in a progressive manner. The same or similar parts between the various embodiments can be referred to each other. Each embodiment focuses on describing the differences from other embodiments.

[0039] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A holographic fault location system for low-voltage power distribution lines, characterized in that, include: Data acquisition module: used to collect real-time, multi-dimensional operational data of low-voltage power distribution lines, construct holographic data for fault diagnosis, and the collection scope covers line electrical parameters, environmental parameters, line physical state parameters and equipment operating state parameters; Data preprocessing module: Connected to the data acquisition module, it is used to remove noise, filter redundancy, and complete the integrity of the acquired raw data; Data calibration module: Built-in multi-dimensional standard calibration model library, preset corresponding calibration parameters and correction algorithms for different types of collected data, and perform accuracy calibration and correction of deviations on pre-processed clean data; Holographic data fusion module: It adopts a three-level progressive holographic fusion strategy to perform cross-dimensional and full-element holographic fusion processing on the calibrated standardized data to construct a holographic dataset that comprehensively depicts the line's operating status; Fault Feature Extraction Module: Constructs a multi-type fault feature template library, and uses a hybrid feature extraction method combining deep learning and traditional signal processing to extract feature data related to various faults from holographic fusion data; Holographic Diagnostic Analysis Module: Integrates an analytical model based on fault mechanism, a machine learning diagnostic model based on deep learning, and a rule-based reasoning model based on expert experience. Through the synergistic linkage of these three models, it performs full-dimensional and multi-perspective holographic diagnosis of low-voltage power distribution line faults. Precise positioning calculation module: It adopts a multi-algorithm fusion positioning strategy, combines the temporal and spatial features of holographic data to complete the positioning calculation, and realizes the precise positioning of the fault point; Diagnostic result verification module: A four-level closed-loop verification mechanism is used to verify the diagnostic conclusions and fault location results in a multi-dimensional and full-link manner.

2. The holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The data acquisition module specifically includes: Multi-type sensing units: collect line electrical parameters, environmental parameters, line physical state parameters and equipment operating status parameters, specifically including: three-phase voltage, three-phase current, zero-sequence current, line active power, reactive power, power factor electrical parameters; temperature, humidity, wind speed, precipitation, ice thickness environmental parameters; line joint temperature, insulator pollution level, conductor sag line physical state parameters; and operating status parameters of distribution transformers, circuit breakers, fuses equipment. Hybrid communication unit: LoRa self-organizing network transmission is used for short distances, and Ethernet is used to access the distribution network backbone for long distances; Clock synchronization unit: integrates Beidou time synchronization module, dynamically calibrates the timestamps of each acquisition terminal through NTP protocol.

3. The holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The data preprocessing module specifically includes: Noise Removal Unit: Identifies abnormal noise data through a multi-threshold screening mechanism, adaptively adjusts the filtering window according to the intensity of data fluctuations, and removes noise caused by electromagnetic interference from power distribution lines; Data normalization unit: By distinguishing the dimensional attributes of different types of data such as electrical, environmental and equipment status data, a hierarchical scaling strategy is used to map the data to the [0,1] interval and record the dimensional conversion mapping relationship; Time-series interpolation completion unit: Based on the historical trend of data changes and similar working conditions data of the same period, a piecewise interpolation strategy is used to complete the missing data; Data quality verification unit: Constructs a three-dimensional verification index system of integrity, consistency and accuracy, performs full verification on preprocessed data, and triggers a backtracking and reprocessing process for unqualified data.

4. The holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The data calibration module specifically includes: Standard calibration model library unit: It has three major categories of calibration models: electrical, environmental and equipment status. It presets calibration reference parameters for different data acquisition terminal models and line operating conditions, and stores historical calibration data and deviation correction data. Scene matching and model selection unit: Extracts the acquisition scene information and acquisition terminal identification information of the preprocessed data, and matches it with the corresponding standard calibration model; Deviation correction unit: Calculates the deviation between the measured data and the standard model reference value in real time, and dynamically adjusts the correction coefficient based on the running time of the acquisition terminal and changes in ambient temperature and humidity to complete deviation compensation; Calibration result verification unit: Constructs dual verification indicators of accuracy and consistency, performs full verification of calibrated data, triggers backtracking and recalibration process for unqualified data, and attaches calibration accuracy rating label to qualified data.

5. A holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The holographic data fusion module specifically includes: Data layer fusion unit: performs time axis alignment and redundancy removal on standardized data from different acquisition nodes of the same type, and generates a continuous data sequence covering the entire link through a sliding window aggregation algorithm; Feature layer fusion unit: Extracts time-domain, frequency-domain and time-frequency-domain features from data of each dimension, establishes cross-dimensional feature associations through feature correlation analysis, and generates a unified feature set; Decision-making fusion unit: It assigns weights and resolves conflicts based on the preliminary state judgment results of data from each dimension, and outputs a comprehensive state assessment result after integrating complementary information. Fusion process control unit: coordinates the collaborative work of the three-level fusion units, records fusion process parameters, and triggers a backtracking and re-fusion mechanism for abnormal fusion data.

6. The holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The fault feature extraction module specifically includes: Fault Feature Template Library Unit: Built-in short circuit, grounding, and overload fault feature subsets, storing typical time-domain sudden change thresholds, frequency-domain distortion intervals, and feature change time sequence patterns for each fault, and associating fault causes with feature mapping relationships; Multi-domain feature mining unit: integrates time domain, frequency domain, and deep abstract feature extraction sub-modules. The time domain sub-module captures the peak value and duration of data mutations, the frequency domain sub-module identifies fault frequency distortion features, and the deep sub-module mines non-explicit correlation features. Feature matching and filtering unit: Employs a feature similarity comparison algorithm to accurately match multi-domain mined features with a template library, generating targeted fault feature vectors; Feature quality control unit: performs integrity and consistency checks on the generated feature vectors, and triggers a backtracking and re-mining process for unqualified features.

7. A holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The holographic diagnostic analysis module specifically includes: Multi-model fusion diagnostic engine unit: integrates three sub-modules: analytical model, machine learning diagnostic model and rule reasoning model. The analytical model constructs a logical judgment link based on the fault mechanism to lock the range of fault types. The machine learning model is equipped with a deep training network to mine hidden correlation information in holographic data. The rule reasoning model integrates the domain expert experience base to form a complementary diagnostic link. Model Coordination Scheduling Unit: Employs a dynamic weight allocation mechanism to adjust the output weights of each model based on the complexity of the fault scenario, integrates differing conclusions through a conflict resolution algorithm, and outputs a unified diagnostic result; Learning and updating unit: Real-time absorption of new fault data and diagnostic experience, and dynamic optimization of model parameters and weight configuration.

8. A holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The precise positioning calculation module specifically includes: Preliminary location calculation unit: Based on the electrical parameters and inherent parameters of the line in the holographic fusion data, a fault impedance analysis link is constructed to obtain the initial distance between the fault point and the acquisition node; Positioning result correction unit: Based on the propagation time and reflection characteristics of the fault traveling wave, calculate the propagation speed and time difference of the traveling wave, perform deviation compensation on the preliminary positioning result, and obtain the corrected accurate distance data; Spatial topology mapping unit: Converts the corrected fault distance into geographic coordinates, associates them with the corresponding line section and equipment identification information, generates distance-coordinate-equipment three-dimensional association data, and accurately maps the fault point space; Complex Scene Adaptation Unit: For complex scenes such as line branches and intersections, dynamically match the corresponding scene positioning algorithm parameters.

9. A holographic fault location system for low-voltage power distribution lines according to claim 1, characterized in that, The diagnostic result verification module specifically includes: Data consistency verification unit: Extract the change curves of holographic fusion data before and after the fault, compare the matching degree between the diagnostic conclusions and the data mutation characteristics and trend anomalies, screen unqualified diagnostic results and mark the deviation type; Model cross-validation unit: By reanalyzing the same fault data through the parsing, machine learning, and rule-based reasoning models of the holographic diagnostic core, the consistency coefficient of the diagnostic conclusions of each model is calculated to verify the stability of the diagnosis. Spatial correlation verification unit: Combining real-time monitoring images of the line, UAV inspection data and equipment operation status feedback information, the fault point location results are verified in the field. Historical Case Verification Unit: The current fault's feature vector, diagnostic conclusion, and location information are compared with the historical fault case database to analyze the correlation between fault type, cause, location, and influencing factors, and to verify the rationality of the diagnostic and location results.

Images