An electric leakage and electricity theft prevention positioning system and device

Abstract

This invention belongs to the field of power system safety monitoring technology, and specifically relates to a leakage and theft prevention and location system and device. The system includes a multimodal information acquisition module, a data preprocessing and fusion module, a feature extraction module, a cause analysis and differentiation module, a precise location module, and a differentiated response module. The multimodal information acquisition module acquires multi-dimensional data on electrical parameters and the environment, and extracts feature sets after preprocessing and fusion. The cause analysis and differentiation module achieves accurate judgment of anomaly type and cause through feature classification and dual threshold matching. The precise location module achieves high-precision location by combining topology data and multi-parameter correction. The differentiated response module provides specific handling measures for leakage and theft triggers. The device is an embedded device that carries all system functions. This invention solves the problems of ambiguous anomaly judgment and low location accuracy in traditional technologies, improves the accuracy and efficiency of anomaly handling, and protects power grid safety and the rights and interests of power supply companies.

Description

Technical Field

[0001] This invention belongs to the field of power system safety monitoring technology, and in particular relates to a leakage current and electricity theft location system and device. Background Technology

[0002] With the expansion of my country's power distribution network and the surge in electricity load, leakage and electricity theft have become prominent hidden dangers restricting the safe and economical operation of the power system. Leakage not only wastes electrical resources but also easily triggers electrical fires and other safety accidents; electricity theft directly leads to revenue losses for power supply companies, and the unauthorized wiring and other operations by thieves can also induce line faults. Statistics show that annual power losses due to leakage and theft amount to billions of yuan, and related safety incidents account for more than 30% of power grid safety incidents. Therefore, the development of efficient leakage and theft detection technologies is an urgent practical need.

[0003] Existing technologies for preventing leakage and locating electricity theft mostly rely on monitoring single electrical parameters such as current and voltage, and judge anomalies through simple methods such as threshold comparison. They lack a systematic analysis mechanism of "multi-source acquisition - feature extraction - cause comparison" and the ability to accurately locate anomalies by combining the topology of the power distribution network.

[0004] The core flaw in existing technology lies in its inability to distinguish the essential characteristics of leakage current and electricity theft, specifically manifested as follows:

[0005] 1. Limited data acquisition dimensions and insufficient data fusion: Data collection relies solely on current transformers to acquire single electrical parameters, failing to utilize multiple types of sensors to collaboratively acquire multimodal information such as insulation resistance, temperature and humidity, and load curves. Furthermore, it lacks standardized cleaning and spatiotemporal fusion processing. For example, when scenarios with similar electrical parameter characteristics, such as decreased line insulation in humid environments and illegal power theft, occur, the false positive rate exceeds 40%, significantly increasing investigation costs.

[0006] 2. Lack of Feature Library Support in Analysis Mechanism: Judgment is based solely on the appearance of abnormal electrical parameters, without a pre-set feature library to support the causes, and multi-dimensional feature extraction is not carried out. For example, both equipment aging and leakage and electricity theft by modified meters manifest as abnormal current, but the former is accompanied by a decrease in insulation resistance, while the latter shows a sudden change in off-peak load. Current technology cannot identify such differences.

[0007] 3. Vague location and inconsistent response: Without establishing a source tracing mechanism based on the power distribution network topology, it is difficult to accurately locate abnormal points. All abnormalities are handled using an "alarm + manual investigation" mode, ignoring the differences in handling different abnormalities, which can easily lead to over-handling or under-response.

[0008] In summary, existing technologies, due to incomplete data acquisition and missing feature extraction, are no longer able to meet the complex power grid monitoring needs. Developing a system that integrates multimodal acquisition, feature library judgment, topology positioning, and differentiated response has become an urgent need in this field. Summary of the Invention

[0009] The purpose of this invention is to provide an anti-leakage and electricity theft location system and device to solve the problem mentioned in the background art that it cannot adapt to the complex power grid monitoring needs.

[0010] To achieve the above objectives, the present invention adopts the following technical solution: a leakage current and electricity theft prevention and positioning system and device, comprising a multimodal information acquisition module, a data preprocessing and fusion module, a feature extraction module, a cause analysis and differentiation module, a precise positioning module and a differentiated response module;

[0011] The cause analysis and differentiation module compares the extracted features with a preset feature library and uses a dual threshold to determine the anomaly type. The first threshold is a feature similarity benchmark value, and features above this threshold are marked as candidate related features. The second threshold is a candidate related feature proportion benchmark value, and features above this proportion are determined to be initially matched with the anomaly type.

[0012] If the proportion of candidate related features of equipment status and environmental features is higher than that of electrical parameters and electricity consumption behavior features, and the feature subset in the matching feature library is the cause of leakage, then it is determined to be leakage; if the proportion of candidate related features of electricity consumption behavior and abnormal electrical parameters is higher, and the feature subset in the matching feature library is the cause of electricity theft, then it is determined to be electricity theft.

[0013] The precise positioning module uses the fused data collected by each monitoring node to inversely deduce the direction of anomaly propagation based on the time difference of the occurrence of abnormal features and the signal attenuation amplitude, and corrects the positioning deviation by combining the line impedance parameters.

[0014] The differentiated response module triggers a response based on the type of anomaly: when there is leakage, it sends a maintenance level notification to the operation and maintenance terminal according to the degree of insulation damage; when there is electricity theft, it automatically intercepts the electricity consumption data during the abnormal period, generates an evidence chain, and pushes it to the inspection terminal.

[0015] Preferably, the multimodal information acquisition module acquires multimodal information in real time through multiple types of sensors deployed at various monitoring nodes of the power distribution network to obtain multi-dimensional raw data;

[0016] The data preprocessing and fusion module cleans, standardizes, and performs multimodal fusion operations on the multidimensional raw data to obtain fused data in a unified format.

[0017] The feature extraction module inputs the fused data into the big data analysis platform to perform multi-dimensional feature extraction, thereby obtaining a multi-dimensional feature set.

[0018] The cause analysis and differentiation module compares and analyzes the multi-dimensional feature set with the preset leakage and electricity theft cause feature library to obtain the judgment result of the abnormal electricity use type and specific cause.

[0019] The precise positioning module combines the judgment result with the power distribution network topology data to perform anomaly tracing and location, thereby obtaining precise positioning information.

[0020] The differentiated response module associates and matches the precise positioning information with the judgment result, triggering the corresponding differentiated response measures.

[0021] Preferably, the multimodal information acquisition module and the data preprocessing and fusion module collaboratively perform information acquisition and processing operations, including:

[0022] The multimodal information acquisition module collects electrical parameter information, monitoring point environmental information, line equipment information, and user electricity consumption behavior information through electrical parameter sensors, environmental sensors, equipment status sensors, and electricity monitoring terminals, and then synchronously associates and stores these various types of information to form multi-dimensional raw data.

[0023] The data preprocessing and fusion module performs outlier removal, noise filtering, and standardization on the multi-dimensional raw data collected and associated by the multi-modal information acquisition module. Then, it establishes information correspondence through timestamp synchronization and spatial index association, and integrates the data into a unified spatiotemporally associated format.

[0024] Preferably, the process of determining the type and specific cause of power consumption anomalies performed by the cause analysis and differentiation module includes:

[0025] Based on the multi-dimensional feature set output by the feature extraction module, the feature set is first classified into electrical parameter abnormality features, environmental correlation features, equipment status features, and electricity consumption behavior features. Then, it is matched and associated with the corresponding feature subsets in the preset leakage and electricity theft cause feature library. Finally, the judgment result of the abnormal electricity consumption type and specific cause is output.

[0026] Preferably, the anomaly tracing and localization performed by the precise localization module includes:

[0027] Guided by the judgment results of the power consumption anomaly type and specific cause output by the cause analysis and differentiation module, and combined with the distribution network topology data containing the distribution area-branch line-user terminal hierarchical topology relationship, the distribution area range is first locked by the abnormal signal strength in the fused data, then the current fluctuation difference of the branch line is compared to narrow it down to the specific line, and finally the abnormal point is determined by matching the unique power consumption characteristics of the user terminal.

[0028] Preferably, the preset feature library is a preset feature library of leakage and electricity theft causes, which includes a dynamic update mechanism. It can associate and store the abnormal features and actual causes of each judgment, and improve the accuracy of subsequent judgments by adding new samples to optimize the matching parameters of feature subsets.

[0029] The present invention provides a device for preventing leakage current and electricity theft, used to perform the method for preventing leakage current and electricity theft implemented by any of the anti-leakage current and electricity theft positioning systems described in Examples 1 to 9.

[0030] Compared with existing technologies, the advantages of this anti-leakage and electricity theft location system and device are as follows:

[0031] 1. Accurate and efficient anomaly detection: Through multimodal data acquisition and feature classification matching mechanism, combined with dual threshold judgment logic, the essential characteristics of leakage and electricity theft can be distinguished, solving the problem of "easily confused abnormal electrical parameters" in traditional technology, and significantly improving the accuracy of anomaly type and cause judgment.

[0032] 2. Improved positioning accuracy at each level: Based on the hierarchical relationship of the power distribution network topology, combined with abnormal signal analysis and multi-node parameter correction, the positioning range is gradually narrowed from the traditional "transformer area level" to the "user terminal level" and even the "specific equipment or line section level", which greatly reduces the difficulty and time cost of operation and maintenance troubleshooting.

[0033] 3. Strong adaptability in response and handling: It has a differentiated response mechanism and the ability to dynamically update the feature library. It pushes maintenance notices according to the risk level of leakage and automatically generates a compliant evidence chain for electricity theft. At the same time, it can optimize parameters and expand feature subsets by adding new samples to adapt to new abnormal scenarios and enhance the long-term adaptability of the system. Attached Figure Description

[0034] Figure 1 This is a schematic diagram of an anti-leakage and electricity theft location system in an embodiment of the present invention. Detailed Implementation

[0035] The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0036] Example 1:

[0037] This invention provides a leakage current theft and location system, referring to... Figure 1 It includes: a multimodal information acquisition module, a data preprocessing and fusion module, a feature extraction module, a causal analysis and differentiation module, a precise positioning module, and a differentiated response module. Each module works together to perform the following operations:

[0038] The multimodal information acquisition module collects multimodal information in real time through various types of sensors deployed at monitoring nodes of the power distribution network, and obtains multi-dimensional raw data.

[0039] The data preprocessing and fusion module cleans, standardizes, and fuses multi-dimensional raw data to obtain fused data in a unified format.

[0040] The feature extraction module will input the fused data into the big data analysis platform to perform multi-dimensional feature extraction, resulting in a multi-dimensional feature set;

[0041] The cause analysis and differentiation module compares and analyzes the multi-dimensional feature set with the preset leakage and electricity theft cause feature library to obtain the judgment results of the abnormal electricity use type and specific cause.

[0042] The precise positioning module combines the judgment results with the power distribution network topology data to trace and locate the source of the anomaly, thereby obtaining precise positioning information.

[0043] The differentiated response module associates and matches precise location information with judgment results, triggering corresponding differentiated response measures to achieve accurate differentiation, location, and handling of leakage and theft of electricity.

[0044] In this embodiment, the multi-modal information acquisition module requires the deployment of various types of sensors according to the power distribution network hierarchy: the main control node of the transformer substation is equipped with a composite sensor integrating electrical parameters and environmental monitoring; dedicated sensors for equipment status are deployed on branch lines; and user terminals are equipped with electricity consumption behavior monitoring terminals, forming a three-level acquisition network of "main-branch-terminal". For example, in old residential areas, in response to the problem of aging lines, additional high-frequency monitoring sensors are installed in sections with weak insulation to ensure that the raw data covers the entire power grid operation scenario.

[0045] In this embodiment, the cleaning operation of the data preprocessing fusion module includes outlier removal and noise filtering. For electrical parameter data, statistical criteria are used to remove instantaneous spikes, and environmental data is smoothed out by filtering algorithms. Standardization processing converts data of different magnitudes to a unified dimension range. Multimodal information fusion uses "station number - acquisition time" as a joint index to establish the spatiotemporal correlation of electrical parameters, environment, and equipment status.

[0046] In this embodiment, the multi-dimensional features of the feature extraction module cover four core features: abnormal electrical parameters (such as current fluctuation coefficient), environmental correlation features (such as the correlation between temperature and humidity and insulation resistance), equipment status features (such as insulation resistance attenuation rate), and electricity consumption behavior features (such as load period deviation). The feature dimensions need to meet the requirement of comprehensively reflecting the power grid operation status.

[0047] In this embodiment, the preset leakage current theft cause feature library is built based on a large number of historical abnormal cases. It contains feature templates of various typical abnormalities. Each type of template contains multiple core feature parameters and threshold ranges, which can provide a reliable basis for comparison and analysis.

[0048] In this embodiment, the distribution network topology data of the precise positioning module is connected to the power grid GIS system, including the physical connection relationship and coordinate information of "distribution transformer - main switch - branch line - user terminal". Combined with abnormal signal strength gradient analysis, the hierarchical locking of the abnormal range is realized.

[0049] In this embodiment, the differentiated response module has two pre-defined handling procedures for differentiated response measures: leakage current triggers maintenance dispatch, which includes location coordinates and suggested solutions, and is pushed to the operation and maintenance terminal; electricity theft triggers evidence fixation, which automatically generates a standardized evidence chain including load curves and pushes it to the inspection terminal.

[0050] The beneficial effects of the above technologies are as follows: through the coordinated operation of six modules, the entire process from data collection to anomaly handling is automated, which solves the problems of single data, ambiguous judgment and inaccurate positioning in traditional monitoring technologies, significantly improves the accuracy of anomaly judgment and positioning accuracy, shortens the operation and maintenance investigation time, and effectively protects the safety of the power grid and the rights and interests of power supply companies.

[0051] Example 2:

[0052] Based on Example 1, a leakage current theft location system includes a multimodal information acquisition module and a data preprocessing and fusion module that collaboratively perform information acquisition and processing operations, comprising:

[0053] The multimodal information acquisition module collects electrical parameter information, monitoring point environmental information, line equipment information, and user electricity consumption behavior information through electrical parameter sensors, environmental sensors, equipment status sensors, and electricity consumption monitoring terminals. After synchronously associating and storing various types of information, it forms multi-dimensional raw data.

[0054] The data preprocessing and fusion module performs outlier removal, noise filtering, and standardization on the multi-dimensional raw data collected and associated by the multi-modal information acquisition module. Then, it establishes information correspondence through timestamp synchronization and spatial index association, and integrates it into a unified format of spatiotemporally associated fused data.

[0055] In this embodiment, the synchronous associated storage of the multimodal information acquisition module is achieved through system clock calibration. The time error of all sensors and terminals is controlled within a very small range, and the acquired data is attached with a three-level identifier of "transformer area - line - terminal" to ensure that the data is traceable.

[0056] Different data collection cycles are set for different devices: core electrical parameters are collected at high frequency to capture instantaneous anomalies; environmental data are collected at medium frequency to balance accuracy and energy consumption; user electricity consumption behavior data is collected in time periods, with the collection frequency increased during peak electricity consumption periods.

[0057] In this embodiment, the electrical parameter sensor is of high precision type to ensure the accuracy of electrical parameter measurement; the environmental sensor supports temperature, humidity and special weather parameter monitoring; the equipment status sensor focuses on core indicators such as insulation resistance and cable temperature; the power monitoring terminal collects single-phase or three-phase load data and supports real-time uploading of load curves.

[0058] In this embodiment, the outlier removal in the data preprocessing and fusion module is based on specific rules for different data types:

[0059] Electrical parameter data is filtered out for extreme values ​​that deviate from the normal range, and equipment status data is filtered out for invalid values ​​that exceed the measurement range;

[0060] Noise filtering employs filtering algorithms adapted to various types of data to process interference signals;

[0061] The standardization process uses a universal standardization algorithm to convert the data into a unified standard format.

[0062] In this embodiment, timestamp synchronization is based on a high-precision clock, and the device time is calibrated periodically; spatial index association adopts a grid indexing method, which divides the power distribution area into equal grids, assigns a unique index value to each grid, and realizes precise binding between data and physical location.

[0063] In a specific application scenario, in a residential area in a rainy region, the multimodal information acquisition module simultaneously collects the current, ambient humidity, insulation resistance and user load curves of the branch lines, and stores them as raw data containing three levels of identifiers and timestamps.

[0064] The data preprocessing and fusion module removes instantaneous spikes in current data and filters high-frequency fluctuations in humidity data. After standardization, it is linked by timestamps and spatial indices to form spatiotemporally unified fused data, providing support for subsequent judgments that "high humidity leads to insulation degradation".

[0065] The beneficial effects of the above technologies are as follows: the multimodal information acquisition module achieves accurate acquisition of various types of information through dedicated sensors, and synchronous associated storage ensures the correlation of the original data; the series of operations of the data preprocessing and fusion module effectively improves data quality, spatiotemporal correlation integration solves the problem of messy multi-source data, and the generated fused data provides high-quality support for subsequent analysis and reduces the error of basic data.

[0066] Example 3:

[0067] Based on Example 1, the process of determining the type and specific cause of power consumption anomalies performed by the cause analysis and differentiation module includes:

[0068] Based on the multi-dimensional feature set output by the feature extraction module, the feature set is first classified into abnormal electrical parameters, environmental correlation features, equipment status features, and electricity consumption behavior features. Then, it is matched and associated with the corresponding feature subsets in the preset leakage and electricity theft cause feature library. Finally, the judgment result of the abnormal electricity consumption type and specific cause is output.

[0069] In this embodiment, the attribute division criteria for feature classification are clear:

[0070] Abnormal electrical parameter characteristics focus on abnormal changes in electrical indicators such as current, voltage, and power, and include multiple characteristics reflecting the electrical state;

[0071] Environmental correlation features focus on the relationship between environmental parameters and line status, and include multiple environmental and equipment correlation features;

[0072] Equipment status characteristics revolve around the performance of the lines and the equipment itself, including multiple characteristics that reflect the health of the equipment;

[0073] Electricity consumption behavior characteristics are based on users' electricity consumption patterns and include multiple features that reflect their electricity consumption habits.

[0074] In this embodiment, the preset feature library of leakage and electricity theft causes is divided into two major feature sub-clusters according to leakage and theft, and each cluster contains a feature subset of various typical anomalies.

[0075] Each feature subset contains exclusive core feature parameters and threshold ranges. For example, the leakage current subset contains a combination of core features such as abnormal insulation-related parameters, high ambient humidity, and stable current.

[0076] In this embodiment, the matching association uses a general similarity algorithm to calculate feature similarity. When the similarity reaches the set high association threshold, it is determined to be a strong match. The medium association threshold range is a match to be verified. The range below the medium association threshold is a weak match. Only high association features participate in subsequent analysis.

[0077] In a specific application scenario, an anomaly was detected in the monitoring area of ​​a commercial complex. After the feature extraction module outputs a multi-dimensional feature set, the cause analysis and differentiation module classifies it into four types of features.

[0078] Subsequently, it was matched with the feature library. The similarity between the device status feature and the leakage category subset reached the high correlation threshold, the similarity between the environmental correlation feature and the subset reached the medium correlation threshold, and the similarity between other features and the electricity theft category subset was all below the medium correlation threshold. The initial screening direction of the cause was leakage. Finally, a clear anomaly type and specific cause judgment result were output.

[0079] The beneficial effects of the above technologies are as follows: feature classification structurates complex feature sets and clarifies the core orientation of various features; similarity matching ensures the accuracy of the correlation, avoids blind correspondence between features and causes, improves the targeting of anomaly cause analysis, and lays the foundation for subsequent accurate judgment.

[0080] Example 4:

[0081] Based on Example 3, the matching association process in determining the type and specific cause of power consumption anomalies employs a dual threshold determination:

[0082] The first threshold is the feature similarity benchmark value. When the similarity between a certain type of feature and the corresponding feature subset in the preset feature library of leakage and electricity theft causes is higher than the threshold, the feature is marked as a candidate associated feature.

[0083] The second threshold is the baseline value of the proportion of associated features. When the proportion of candidate associated features to the total number of features of this type is higher than the second threshold, it is determined that this type of feature is initially matched with the corresponding anomaly type.

[0084] In this embodiment, the dual thresholds are determined through training with a large amount of historical data, and the default setting requires a balance between judgment accuracy and sensitivity.

[0085] The threshold can be dynamically adjusted for different application scenarios. For example, in industrial areas where frequent equipment starts and stops cause large fluctuations in electrical parameters, the first threshold can be appropriately increased to improve the rigor of the judgment; in residential areas where the power supply is stable, the first threshold can be appropriately decreased to improve the detection sensitivity.

[0086] In this embodiment, the candidate association feature labeling process is as follows: for each feature item in a certain type of feature, calculate the similarity with the template feature of the corresponding component set. If the similarity of a single feature item is higher than the first threshold, then the feature item is labeled as a candidate association feature.

[0087] In this embodiment, the proportion of associated features is calculated as the ratio of the number of candidate associated feature items to the total number of feature items of that type. When the proportion is higher than a second threshold, it is determined that the feature type initially matches the corresponding anomaly type.

[0088] In a specific application scenario, an anomaly occurs in the commercial complex area of ​​Implementation Example 3, and dual thresholds are set according to the scenario requirements.

[0089] Some features in the equipment status characteristics are candidate related features, but their proportion is below the second threshold; the proportion of candidate related features in the environmental related features is higher than the second threshold; the proportion of candidate related features in electrical parameters and electricity consumption behavior characteristics is lower than the second threshold. Therefore, it is determined that the environmental related features are initially matched with leakage current anomaly types, and this judgment direction is strengthened by the high similarity of the equipment status characteristics.

[0090] The beneficial effects of the above technology are as follows: the dual threshold judgment forms a verification logic of "single feature reliability - multi feature collaboration". The first threshold filters out interference from accidental similarity, and the second threshold avoids misjudgment of a single feature, which significantly improves the reliability of the initial matching results, reduces the misjudgment rate, and provides more reliable intermediate results for subsequent anomaly type differentiation.

[0091] Example 5:

[0092] Based on Example 4, the abnormality type differentiation logic in the process of determining the type and specific cause of power consumption abnormalities includes:

[0093] If the proportion of candidate associated features of equipment status features and environmental associated features is higher than that of electrical parameters and electricity consumption behavior features, and the feature subset in the preset leakage and electricity theft cause feature library is the cause of leakage, then it is determined to be leakage.

[0094] If the proportion of candidate related features of electricity consumption behavior characteristics and abnormal electrical parameters is higher, and the feature subset in the preset leakage and electricity theft cause feature library matches the cause of electricity theft, then it is determined to be electricity theft.

[0095] In this embodiment, the candidate association ratio comparison is calculated using a weighted ratio. Since the association weights of different features and anomaly types are different: the association features of equipment status and environment have a higher weight for leakage current, and the association features of electricity consumption behavior and abnormal electrical parameters have a higher weight for electricity theft, ensuring that the influence of core features is more significant.

[0096] In this embodiment, the core features of the feature subset corresponding to the cause of leakage are clear, focusing on the device's own state and environmental influences; the core features of the feature subset corresponding to the cause of electricity theft focus on abnormal electricity use caused by user intervention, and the feature orientations of the two types of subsets are significantly different.

[0097] In a specific application scenario (leakage determination), following the scenario of Example 4, the weighted sum of the characteristics related to the device status and the environment is significantly higher than that of the characteristics related to electrical parameters and electricity consumption behavior, and the matching subset is the cause of leakage. Finally, it is determined to be leakage, and the specific cause is clearly related to environmental factors.

[0098] In a specific application scenario (electricity theft determination), an anomaly was detected in a residential transformer area. After feature classification, the weighted sum of the candidate association ratios of electricity consumption behavior and abnormal electrical parameters was significantly higher than that of other features. The matching subset was the cause of electricity theft, and the final determination was electricity theft, with the specific cause pointing to the user's illegal electricity consumption behavior.

[0099] The beneficial effects of the above technologies are as follows: a distinguishing logic is established based on the essential differences between leakage and electricity theft; the influence of core features is strengthened by weighted ratio comparison; and accurate distinction is achieved by combining feature subset matching types. This solves the problem of easy confusion due to abnormal electrical parameters in traditional technologies and provides an accurate basis for subsequent differentiated handling.

[0100] Example 6:

[0101] Based on Example 3, the anomaly tracing and localization performed by the precise localization module includes:

[0102] Guided by the judgment results of the power consumption anomaly type and specific cause output by the cause analysis module, and combined with the distribution network topology data containing the distribution area-branch line-user terminal hierarchical topology relationship, the distribution area range is first locked by the abnormal signal strength in the fused data, then the current fluctuation difference of the branch line is compared to narrow it down to the specific line, and finally the anomaly point is determined by matching the unique power consumption characteristics of the user terminal.

[0103] In this embodiment, the power distribution network topology data is a digital model, containing a three-level hierarchical relationship and core parameters:

[0104] The distribution area level includes information on distribution transformers and main switches;

[0105] Branch line level includes line attributes and parameters;

[0106] The user terminal level includes meter boxes and user information, and the data is updated in real time through the power grid GIS system.

[0107] In this embodiment, the intensity of abnormal signals is quantitatively represented by the degree of characteristic anomalies in the fused data. An anomaly index is used for calculation, and the higher the index, the more severe the anomaly. Abnormal stations are identified by comparing the anomaly indices of different stations.

[0108] In this embodiment, the difference in current fluctuation is calculated using statistical coefficients. If the fluctuation coefficient exceeds the normal range, it is determined to be abnormal. By comparing the fluctuation coefficients of each branch line within the same transformer area, the abnormal branch line is identified.

[0109] In this embodiment, the unique electricity consumption characteristics of a user terminal include the user's load curve characteristics, electricity consumption patterns and preferences. Each user terminal corresponds to a unique feature vector, and anomaly locations are determined through similarity matching.

[0110] In a specific application scenario, an anomaly was detected in the power grid of a certain area, and the cause analysis result was "leakage and insulation-related issues". The precise positioning module retrieved topology data, calculated the anomaly index of each transformer substation to pinpoint the target substation; compared the current fluctuation coefficient of the branch lines within the substation to locate the abnormal branch; extracted the unique power consumption characteristics of the user terminal on this branch, and determined the anomaly location to be a certain user's meter box related line through similarity matching.

[0111] The beneficial effects of the above technologies are as follows: relying on the three-level topology relationship to achieve hierarchical locking of the abnormal range, the positioning logic from the transformer area to the user terminal is clear, the joint application of multi-dimensional features ensures accurate positioning, and the positioning range is narrowed from the traditional "transformer area level" to the "user terminal level", which significantly improves the troubleshooting efficiency of operation and maintenance personnel.

[0112] Example 7:

[0113] Based on Example 6, anomaly tracing and localization also includes:

[0114] By analyzing the time difference of abnormal features and the amplitude of signal attenuation in the fused data collected from each monitoring node, the propagation direction of the abnormal source is deduced in reverse. Combined with the line impedance parameters contained in the power distribution network topology data, the positioning deviation is corrected, making the positioning accuracy accurate to the user's electricity meter or specific faulty equipment.

[0115] In this embodiment, monitoring nodes are deployed at uniform intervals on the branch lines, each node is uniquely numbered, and the occurrence time and signal strength of abnormal features are collected synchronously. The time recording accuracy reaches the millisecond level, providing support for calculating the time difference.

[0116] In this embodiment, the time difference in the appearance of abnormal features is the time difference between the first detection of the abnormality by different monitoring nodes. Combined with the signal propagation speed, the approximate range of the abnormality source is initially determined.

[0117] In this embodiment, the signal attenuation amplitude is the intensity attenuation value of the abnormal signal propagating from the source to each node. An attenuation model is established in combination with the line impedance parameters to calculate the distance from the abnormal source to each node.

[0118] In this embodiment, the positioning deviation correction adopts a multi-node joint positioning method, which uses the positions of multiple monitoring nodes as a reference and combines the calculated distance values ​​to construct a positioning model, determine the corrected abnormal points, and further improve the positioning accuracy.

[0119] In a specific application scenario, following the abnormal branch line of Example 6, the time difference of abnormal occurrence and signal attenuation amplitude of multiple monitoring nodes are collected. Combined with the line impedance parameters, the distance from the abnormal source to each node is calculated. Through multi-node joint positioning correction, the abnormal point is determined to be a specific location of the incoming line of a certain user's meter box. The positioning accuracy meets the requirements of targeted handling.

[0120] The beneficial effects of the above technologies are as follows: by combining the analysis of time difference, signal attenuation amplitude and line impedance, the positioning deviation can be accurately corrected, and the positioning accuracy can be improved from "user terminal level" to "equipment or line section level", which fully meets the needs of operation and maintenance personnel for targeted handling, avoids ineffective excavation and repeated maintenance, and reduces operation and maintenance costs.

[0121] Example 8:

[0122] Based on Example 1, the preset leakage current theft cause feature library includes a dynamic update mechanism, which can associate and store the abnormal features and actual causes of each judgment, and improve the accuracy of subsequent judgments by adding new samples to optimize the matching parameters of feature subsets.

[0123] In this embodiment, the associated stored content forms a complete data chain, from the original feature data, the processed feature set, the judgment result, the on-site verification of the cause to the handling feedback. Each data chain is assigned a unique identifier and stored in categories according to "time - anomaly type" for easy retrieval and analysis.

[0124] In this embodiment, the newly added sample optimization matching parameters adopt a machine learning algorithm with the judgment accuracy as the objective function. The parameter optimization is triggered according to a set period or sample number, and the template threshold and feature weight parameters of the feature subset are adjusted to make the matching parameters more in line with the actual application scenario.

[0125] In this embodiment, dynamic updates include incremental updates and extended updates: incremental updates target existing anomaly types and optimize the parameters of existing feature subsets; extended updates target new anomaly types, automatically identify their unique features, establish new feature subsets, and supplement them to the feature library.

[0126] In a specific application scenario, a new type of electricity theft appeared in an industrial park. When the system first monitored it, it was identified as an unknown anomaly. After on-site verification, the unique characteristics of the anomaly were associated with the actual cause and stored as a new sample. The system optimized the feature library parameters and added corresponding feature subsets. When similar characteristics are monitored again in the future, accurate judgment can be achieved.

[0127] The beneficial effects of the above technologies are as follows: the dynamic update mechanism frees the feature library from the limitations of rigidity, incremental updates improve the judgment accuracy of existing types, extended updates adapt to new anomalies, and the system's judgment accuracy gradually improves with the use of time, thus enhancing the system's long-term adaptability and vitality.

[0128] Example 9:

[0129] Based on Example 1, the differentiated response module triggers a response according to the judgment results of the power consumption anomaly type and specific cause output by the cause analysis module:

[0130] When leakage occurs, a maintenance level notification is sent to the maintenance personnel's terminal according to the degree of insulation damage. When electricity theft occurs, the system automatically intercepts electricity consumption data during abnormal periods to generate an evidence chain, which is then simultaneously pushed to the investigation personnel's processing terminal.

[0131] In this embodiment, the maintenance level of leakage current is divided into multiple levels according to the degree of insulation damage. Different levels correspond to different response times and handling priorities. The maintenance notice includes the level, location coordinates, cause, recommended tools, and safety precautions.

[0132] In this embodiment, the evidence chain for electricity theft must comply with the electricity inspection and evidence collection standards, including basic information, data evidence and supporting information, automatically generate standardized format files, and support secure transmission and storage.

[0133] In this embodiment, the information push adopts a dual push mode of "terminal + platform": the mobile terminals of operation and maintenance personnel or inspectors receive real-time message reminders, and the power operation and maintenance platform synchronously stores the notification or evidence chain, supports status tracking, and ensures a closed-loop response.

[0134] In a specific application scenario (leakage response), a branch line in a certain community has a leakage current. Based on the degree of insulation damage, it is determined to be a medium maintenance level. The differentiated response module generates a maintenance notification containing complete information and pushes it to the terminal of the operation and maintenance personnel. The platform tracks the status of the notification processing.

[0135] In a specific application scenario (electricity theft response), if a user engages in illegal electricity use, the system automatically intercepts electricity usage data during abnormal periods, generates a standardized evidence chain containing user information and data evidence, pushes it to the inspector's terminal, and simultaneously uploads it to the inspection platform for record-keeping.

[0136] The beneficial effects of the above technologies are as follows: the differentiated response mechanism enables classified management of "rapid repair for leakage current and accurate evidence collection for electricity theft", the maintenance level classification ensures reasonable allocation of resources, the standardized evidence chain provides strong support for the investigation and handling of electricity theft, the dual push mode ensures a closed-loop response, and improves the efficiency of operation and maintenance and the success rate of the investigation and handling of electricity theft.

[0137] Example 10:

[0138] The present invention provides a device for preventing leakage current and electricity theft, used to perform the method for preventing leakage current and electricity theft implemented by any of the anti-leakage current and electricity theft positioning systems in Examples 1 to 9.

[0139] The beneficial effects of the above technologies are as follows: through the coordinated operation of six modules, the entire process from data collection to anomaly handling is automated, solving the problems of "single data, ambiguous judgment, and inaccurate positioning" in traditional monitoring technologies, significantly improving the accuracy of anomaly judgment and positioning, shortening the operation and maintenance investigation time, and effectively protecting the safety of the power grid and the rights and interests of power supply companies.

[0140] 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 spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A leakage current theft location system, characterized in that, include: The system includes a multimodal information acquisition module, a data preprocessing and fusion module, a feature extraction module, a cause analysis and differentiation module, a precise positioning module, and a differentiated response module. The cause analysis and differentiation module compares the extracted features with a preset feature library and uses a dual threshold to determine the anomaly type. The first threshold is a feature similarity benchmark value, and features above this threshold are marked as candidate related features. The second threshold is a candidate related feature proportion benchmark value, and features above this second threshold are determined to be initially matched with the anomaly type. If the proportion of candidate related features of equipment status and environmental features is higher than that of electrical parameters and electricity consumption behavior features, and the feature subset in the matching feature library is the cause of leakage, then it is determined to be leakage; if the proportion of candidate related features of electricity consumption behavior and abnormal electrical parameters is higher, and the feature subset in the matching feature library is the cause of electricity theft, then it is determined to be electricity theft. The precise positioning module uses the fused data collected by each monitoring node to inversely deduce the direction of anomaly propagation based on the time difference of the occurrence of abnormal features and the signal attenuation amplitude, and corrects the positioning deviation by combining the line impedance parameters. The differentiated response module triggers a response based on the type of anomaly: when there is leakage, it sends a maintenance level notification to the operation and maintenance terminal according to the degree of insulation damage; when there is electricity theft, it automatically intercepts the electricity consumption data during the abnormal period, generates an evidence chain, and pushes it to the inspection terminal.

2. The anti-leakage electricity theft positioning system according to claim 1, characterized in that, The multimodal information acquisition module collects multimodal information in real time through various types of sensors deployed at each monitoring node of the power distribution network, and obtains multi-dimensional raw data. The data preprocessing and fusion module cleans, standardizes, and performs multimodal fusion operations on the multidimensional raw data to obtain fused data in a unified format. The feature extraction module inputs the fused data into the big data analysis platform to perform multi-dimensional feature extraction, thereby obtaining a multi-dimensional feature set. The cause analysis and differentiation module compares and analyzes the multi-dimensional feature set with the preset leakage and electricity theft cause feature library to obtain the judgment result of the abnormal electricity use type and specific cause. The precise positioning module combines the judgment result with the power distribution network topology data to perform anomaly tracing and location, thereby obtaining precise positioning information. The differentiated response module associates and matches the precise positioning information with the judgment result, triggering the corresponding differentiated response measures.

3. The anti-leakage electricity theft positioning system according to claim 2, characterized in that, The multimodal information acquisition module and the data preprocessing and fusion module work together to perform information acquisition and processing operations, including: The multimodal information acquisition module collects electrical parameter information, monitoring point environmental information, line equipment information, and user electricity consumption behavior information through electrical parameter sensors, environmental sensors, equipment status sensors, and electricity monitoring terminals, and then synchronously associates and stores these various types of information to form multi-dimensional raw data. The data preprocessing and fusion module performs outlier removal, noise filtering, and standardization on the multi-dimensional raw data collected and associated by the multi-modal information acquisition module. Then, it establishes information correspondence through timestamp synchronization and spatial index association, and integrates the data into a unified spatiotemporally associated format.

4. The anti-leakage electricity theft positioning system according to claim 2, characterized in that, The process of determining the type and specific cause of power consumption anomalies by the cause analysis and differentiation module includes: Based on the multi-dimensional feature set output by the feature extraction module, the feature set is first classified into electrical parameter abnormality features, environmental correlation features, equipment status features, and electricity consumption behavior features. Then, it is matched and associated with the corresponding feature subsets in the preset leakage and electricity theft cause feature library. Finally, the judgment result of the abnormal electricity consumption type and specific cause is output.

5. The anti-leakage electricity theft positioning system according to claim 4, characterized in that, The anomaly tracing and localization performed by the precise localization module includes: Guided by the judgment results of the power consumption anomaly type and specific cause output by the cause analysis and differentiation module, and combined with the distribution network topology data containing the transformer area-branch line-user terminal hierarchical topology relationship, the transformer area range is first locked by the abnormal signal strength in the fused data, then the current fluctuation difference of the branch line is compared to narrow it down to the specific line, and finally the abnormal point is determined by matching the unique power consumption characteristics of the user terminal.

6. The anti-leakage electricity theft positioning system according to claim 1, characterized in that, The preset feature library is a preset feature library of leakage and electricity theft causes. It includes a dynamic update mechanism that can associate and store the abnormal features and actual causes of each judgment. By adding new samples, the matching parameters of the feature subset are optimized to improve the accuracy of subsequent judgments.

7. A device for preventing leakage and electricity theft, characterized in that, The method for preventing leakage current and electricity theft, implemented by the anti-leakage electricity theft location system according to any one of claims 1 to 6.

Images