A fault identification method and system based on a smart metering switch and a medium

Abstract

This invention relates to the field of smart metering switch technology, specifically providing a fault identification method, system, and medium based on smart metering switches. The method includes: acquiring multi-source data; preprocessing the multi-source data to output a standardized feature dataset; the multi-source data includes electrical data collected by the smart metering switch and meter data from connected smart meters; using the standardized feature dataset as input, triggering an edge computing model through anomaly feature matching, and outputting preliminary fault results; acquiring topology identification signals from connected smart meters according to a preset period, and obtaining dynamic topology relationships based on the topology identification signals; verifying the dynamic topology relationships and preliminary fault results to generate fault assessment results. An edge computing model is deployed locally on the smart metering switch, and the edge computing model is intelligently scheduled for different faults. Through multi-model cross-validation and dynamic topology verification, the fault type and fault source are accurately identified, effectively reducing the probability of misjudgment and missed judgment.

Description

Technical Field

[0001] This invention belongs to the field of intelligent measurement switch technology, and particularly relates to a fault identification method, system and medium based on intelligent measurement switches. Background Technology

[0002] With the upgrading of low-voltage distribution networks towards intelligence, smart metering switches, as core sensing nodes on the edge side, have gradually integrated AC sampling, communication relay, and lightweight edge computing capabilities. Connected to smart meters, each equipped with a topology signal generation module, they can send topology identification signals in real time. The metering switches also feature topology recognition units, laying the foundation for multi-source fusion analysis and dynamic topology identification.

[0003] Currently, the fault assessment model for power distribution networks has the following problems: First, it mainly relies on a single fixed threshold or simple logic threshold, which cannot accurately distinguish various complex faults and the location of the fault. Secondly, the existing model structure is highly homogeneous and has limited generalization ability. It cannot adapt to the differences in load and line in different areas, making it difficult to identify hidden faults such as light load and hidden zero-out. The traditional centralized analysis mode requires uploading the measurement data to the main station for calculation. The delay or interruption of the communication link will directly cause the analysis function to fail, making it impossible to achieve local rapid isolation and recovery of faults. The cross-collection data from the measuring switches and the time-series data from the meters are stored and processed independently, lacking a joint modeling and multi-model hierarchical collaborative judgment architecture, making it difficult to achieve integrated and accurate fault location. At the same time, the meter files are statically configured and cannot dynamically reflect the actual topology connection relationship, resulting in the inability to accurately correlate and effectively integrate multi-source data, including topology signals, load curves, and event records, further weakening the overall performance of fault judgment. Summary of the Invention

[0004] To address the problems existing in the prior art, the present invention provides a fault identification method based on an intelligent measuring switch, comprising the following steps: Step S1: Acquire multi-source data, preprocess the multi-source data, and output a standardized feature dataset; the multi-source data includes electrical data collected by the smart metering switch and meter data received from the connected smart meters. Step S2: Using the standardized feature dataset as input, the edge computing model is triggered by anomaly feature matching to output preliminary fault results, which include fault type, fault location, and confidence level. Step S3: Obtain topology identification signals from the connected smart meter according to a preset cycle, and obtain the dynamic topology relationship between the smart meter and the smart meter based on the topology identification signals; Step S4: Verify the dynamic topology relationship and the preliminary fault results to generate fault assessment results.

[0005] Specifically, the preprocessing in step S1 includes: The multi-source data is aligned to form a time-series synchronized dataset. The timing synchronization dataset is processed using a sliding window filtering algorithm to remove instantaneous interference values. For different types of faults, corresponding fault feature parameters are extracted, and the fault feature parameters are normalized to output a standardized feature dataset.

[0006] Preferably, the edge computing model includes at least a transient mutation calculation model, a time-series correlation analysis model, and a power quality anomaly identification model; the step of triggering the edge computing model through anomaly feature matching specifically includes: Based on the standardized feature dataset, when an abnormal feature matching any edge computing model is detected, the matching edge computing model outputs the preliminary fault result.

[0007] Based on the above scheme, the transient fault calculation model identifies transient faults based on transient characteristic parameters, which at least include voltage surge, voltage drop, transient duration, and instantaneous rate of change of voltage or current; the identification of transient faults specifically includes: The mutation feature parameters are compared with the joint threshold of mutation features to match the electrical characteristics of transient faults, thereby obtaining the predicted transient faults; Simultaneously, the protection action of the intelligent measurement switch is acquired, and the protection action is cross-validated with the predicted instantaneous fault to eliminate false faults. The preliminary result of the instantaneous fault, including fault type and sudden change characteristic parameters, is output.

[0008] Based on the above scheme, the time-series correlation analysis model identifies power outage faults based on power outage feature parameters, which include at least the power outage timestamp of the meter, the characteristics of the voltage loss event cluster, and the voltage status of the switch bus.

[0009] Based on the above scheme, the power quality anomaly identification model identifies power quality faults based on power quality characteristic parameters, which include at least three-phase voltage, phase sequence characteristics, and neutral current; the power quality anomaly identification model calculates phase data and three-phase imbalance based on power quality characteristic parameters to identify phase loss and phase failure power quality faults.

[0010] Based on the above scheme, the edge computing model also includes a topology linkage correction model. The topology linkage correction model acquires topology identification signals at a preset period, generates dynamic topology relationships and updates the transformer area topology map, and maintains meter files based on the transformer area topology map. The meter files include at least the meter number, meter location and communication link status.

[0011] Based on the above scheme, step S4 specifically includes: Obtain preliminary fault results of the edge computing model, perform spatial verification based on the dynamic topology relationship, compare the preliminary fault results that pass the verification, and output the fault judgment result with the highest confidence. The fault judgment result includes at least fault type, fault location and confidence.

[0012] On the other hand, the present invention also provides a fault identification system based on an intelligent measuring switch, used to implement any of the fault identification methods described above. The system is deployed on the intelligent measuring switch and includes a multi-source data acquisition module, a data preprocessing module, and an edge computing module; wherein: The multi-source data acquisition module is used to acquire electrical data through high-frequency electrical signals, and to receive meter data from the connected smart meters. The data preprocessing module is used to preprocess the multi-source data and output a standardized feature dataset; The edge computing module includes a fault identification unit, a topology identification unit, and a fusion judgment unit. The fault identification unit is used to take the standardized feature dataset as input, trigger the edge computing model through abnormal feature matching, and output preliminary fault results. The topology identification unit has a built-in topology identification module, which is used to obtain topology identification signals from the smart meter below according to a preset period, and obtain the dynamic topology relationship between the smart meter and the smart meter based on the topology identification signals. The fusion judgment unit is used to verify the dynamic topology relationship and the preliminary fault results, and generate fault judgment results.

[0013] The present invention also provides a computer-readable storage medium having a computer program, which, when executed by a processor, implements the steps of the fault identification method based on an intelligent measuring switch as described above.

[0014] Compared with the prior art, the present invention has the following beneficial effects: 1. Deploy an edge computing model locally on the smart measurement switch, and intelligently schedule the edge computing model for different faults. Through multi-model cross-validation and dynamic topology verification, accurately identify the fault type and fault source, effectively solve the problem of easy misjudgment by traditional single threshold, and effectively reduce the probability of misjudgment and missed judgment. 2. Data processing and fault identification are completed locally at the measurement switch, eliminating the need to upload raw data to the master station, thus reducing reliance on the master station and improving power supply reliability; at the same time, emergency faults can trigger local tripping and isolation in a very short time, meeting the requirements for rapid on-site handling of distribution networks; 3. The topology linkage correction model activates the meter topology signal generation module, dynamically identifies the affiliation of the downstream meters and updates the meter files in real time. It cross-compares the real-time topology relationship with the fault calculation model results to accurately locate the fault section, thus solving the problem of data not being effectively utilized due to fixed files. 4. Each edge computing model adopts a lightweight design, supports remote upgrades and parameter fine-tuning, and can flexibly adjust model parameters or add dedicated models according to the load characteristics and line conditions of different distribution areas, thereby improving the system's adaptability to different scenarios. Attached Figure Description

[0015] Figure 1 This is an overall flowchart of the fault identification method of the present invention; Figure 2 This is a flowchart illustrating the fault identification method of the present invention. Figure 3 This is a schematic diagram of the transformer area structure of the present invention; Figure 4 This is a schematic diagram of the fault identification system of the present invention. Detailed Implementation

[0016] The invention will be further described below with reference to specific embodiments.

[0017] like Figure 1 and Figure 2 As shown, the present invention provides a fault identification method based on an intelligent measuring switch, comprising the following steps: Step S1: Obtain multi-source data, preprocess the multi-source data, and output a standardized feature dataset; First, the smart metering switch collects electrical data at high frequency through integrated current transformers and temperature sensors, including three-phase voltage, current, body temperature, protection actions, and timestamps. At the same time, the smart metering switch receives meter data from the connected smart meters in real time through HPLC or RS485 communication, including power outage records, time of loss / restoration of voltage, voltage anomaly events, individual household load time sequence data, event sequences, and topology identification signals. Furthermore, the multi-source data undergoes preprocessing, including: Data from multiple sources is aligned based on timestamps to form a time-series synchronized dataset; The time-series synchronized dataset is processed by a sliding window filtering algorithm to remove instantaneous interference values. Specifically, this embodiment uses a sliding window with dynamically adjusted window length, identifies outliers based on the 3σ criterion, marks continuous abnormal data segments, and generates a removal log.

[0018] For different types of faults, corresponding fault feature parameters are extracted, specifically: Based on the dataset after removing interference, for the instantaneous mutation calculation model, mutation feature parameters such as current surge, voltage drop, and mutation duration are calculated; for the time series correlation analysis model, outage feature parameters such as input meter outage timestamp, undervoltage event cluster features, and switch bus voltage status are extracted; for the power quality anomaly identification model, power quality feature parameters such as three-phase voltage, phase sequence features, and neutral current are calculated.

[0019] Furthermore, the above fault feature parameters are subjected to standard normalization processing in the [0, 1] interval to output a standardized feature dataset.

[0020] Step S2: Take the standardized feature dataset as input, trigger the edge computing model through anomaly feature matching, and output the preliminary fault results, which include fault type, fault location and confidence level. Among them, the edge computing model includes at least a transient mutation computing model, a time-series correlation judgment model, and a power quality anomaly identification model; Among them, the edge computing model is triggered by anomaly feature matching: based on a standardized feature dataset, when an anomaly feature that matches any edge computing model is detected, the matched edge computing model outputs the preliminary fault result.

[0021] According to this embodiment, the transient mutation calculation model identifies transient faults based on mutation feature parameters. The method for identifying transient faults is as follows: The mutation feature parameters are compared with the joint threshold of mutation features. The joint threshold of multidimensional features includes current surge threshold, voltage drop threshold, mutation duration threshold and electrical quantity change rate threshold. By comparing the thresholds, the electrical characteristics of transient faults are matched to obtain the predicted transient faults. Simultaneously, the protection action of the intelligent measurement switch is acquired, and the protection action is cross-validated with the predicted instantaneous fault. False faults where the intelligent measurement switch does not act are eliminated, and the preliminary instantaneous fault result containing fault type and sudden change characteristic parameters is output.

[0022] The transient change calculation model in this embodiment accurately matches the inherent electrical characteristics of various transient faults by using multi-dimensional features and threshold criteria. At the same time, it is verified by combining the protection action of the intelligent measurement switch to effectively eliminate false faults caused by transient fluctuations, communication interference, etc.

[0023] According to this embodiment, the time-series correlation analysis model identifies power outage faults based on power outage feature parameters. Through event clustering, time difference comparison, and statistics of the number of out-of-connection households, it determines the type of power outage fault. Specifically, it determines whether it is a complete outage of the main line, a branch-wide outage, or a single-household meter failure based on the number of out-of-connection meters and the voltage of the switch bus. It also determines whether it is an intermittent power outage based on the time difference of the outage. The model outputs the fault type and the outage range.

[0024] According to this embodiment, the power quality anomaly identification model identifies power quality faults based on power quality characteristic parameters. It inputs the three-phase voltage RMS sequence, voltage sampling sequence, and neutral current, and identifies phase loss and phase disconnection power quality faults by calculating phase-by-phase data and three-phase unbalance. Specifically, it determines whether a phase is missing by comparing the voltages of each phase; it determines whether the neutral line is broken based on the neutral current and three-phase voltage; it calculates the total harmonic distortion rate of the voltage sampling sequence and compares it with the harmonic distortion threshold to determine whether harmonics exceed the limit; it calculates the three-phase average voltage to determine whether voltage drift has occurred; and it outputs the fault type and affected area, such as phase loss, with the affected area being phase A.

[0025] Each edge computing model can analyze and determine the fault location based on the fault type and input data. The fault location includes the fault in the measuring switch body, the meter body, or the circuit between the measuring switch and the meter.

[0026] Each edge computing model internally predefines several key feature parameters related to its corresponding fault type, the ideal value range for each feature parameter, and the corresponding weight coefficients. When the model is triggered, it extracts the corresponding feature subset from the standardized feature dataset, compares the measured values ​​of each feature with the predetermined ideal range, calculates the matching degree of each feature (value range 0~1, the closer to 1, the better it matches expectations), and then performs a weighted average of the matching degrees of each feature. The result is the confidence score output by the model. The confidence score characterizes the degree of coupling between the current fault feature and the fault type corresponding to the model. The higher the confidence score, the better the fault feature matches the inherent electrical characteristics of that type of fault.

[0027] Step S3: Obtain topology identification signals from the connected smart meters according to a preset cycle, and obtain the dynamic topology relationship between the smart metering switch and the smart meters based on the topology identification signals; According to this embodiment, the edge computing model also includes a topology linkage correction model. The topology linkage correction model has a built-in topology identification module, which is communicatively connected to the topology signal generation module of the connected smart meter. The smart metering switch activates the topology signal generation module at a preset cycle, triggering the topology signal generation module to inject a topology identification signal. The topology identification module receives the topology identification signal, dynamically identifies the dynamic topology relationship and updates the transformer area topology map. Based on the updated, removed or relocated meter information according to the topology identification signal, the module synchronously identifies and maintains the meter files connected to the smart metering switch. The meter files include the meter number, meter location and communication link status, etc.

[0028] Step S4 involves verifying the dynamic topology relationships and preliminary fault results to generate a fault assessment result. This specifically includes: The system obtains preliminary fault results from the edge computing model, performs spatial consistency verification on the fault location based on the dynamic topology relationship, and removes the model's preliminary results if they are inconsistent with the dynamic topology relationship. The system then compares the preliminary fault results that pass the verification and outputs the fault assessment result with the highest confidence level. The fault assessment result includes fault type, fault location, confidence level, and fault occurrence time.

[0029] Based on the fault assessment results, if an emergency fault such as a short circuit occurs, the intelligent measurement switch will trip locally to quickly isolate the faulty section and prevent the fault from escalating. If a general fault such as a zero-loss, phase loss, or branch power outage occurs, an alarm will be triggered and the fault event will be recorded. At the same time, the fault assessment results will be stored locally and encrypted and uploaded to the intelligent fusion terminal via the HPLC communication module. The intelligent fusion terminal will then report to the main station for maintenance personnel to view and analyze.

[0030] The edge computing model of the present invention supports remote upgrades, and those skilled in the art can fine-tune the model parameters based on actual operating conditions. In other embodiments, those skilled in the art can add other edge computing models according to the load characteristics of different transformer areas and line operating conditions to accurately identify faults and improve the adaptability of the system to different scenarios.

[0031] Based on the same technical concept, this invention provides a fault identification system based on an intelligent measuring switch, used to implement the fault identification method described above, such as... Figure 3 and Figure 4 As shown, the system is deployed on an intelligent measurement switch and includes a multi-source data acquisition module, a data preprocessing module, an edge computing module, and a fault output module; wherein: The multi-source data acquisition module is used to acquire electrical data through high-frequency electrical signals, as well as to receive meter data from connected smart meters. The data preprocessing module is used to preprocess multi-source data and output a standardized feature dataset; The edge computing module includes a fault identification unit, a topology identification unit, and a fusion judgment unit. The fault identification unit takes a standardized feature dataset as input, triggers the edge computing model through abnormal feature matching, and outputs preliminary fault results. The fault identification unit includes at least a transient mutation calculation model, a time-series correlation judgment model, and a power quality anomaly identification model. The topology identification unit obtains the topology identification signal from the smart meter below according to a preset period, and obtains the dynamic topology relationship between the smart meter and the smart meter based on the topology identification signal; The fusion judgment unit is used to verify based on dynamic topology relationships and preliminary fault results, and generate fault assessment results.

[0032] Furthermore, the fault identification method based on an intelligent measuring switch according to the present invention can be recorded in a computer-readable recording medium. Specifically, according to the present invention, a computer-readable recording medium storing computer-executable instructions can be provided, which, when executed by a processor, causes the processor to execute the fault identification method based on an intelligent measuring switch as described above.

[0033] It should be noted that the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present invention. In this regard, each block in a flowchart or block diagram may represent a module, program segment, or portion of code containing at least one executable instruction for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0034] In general, various exemplary embodiments of the present invention can be implemented in hardware or dedicated circuitry, software, firmware, logic, or any combination thereof. Some aspects can be implemented in hardware, while others can be implemented in firmware or software that can be executed by a controller, microprocessor, or other computing device. When aspects of embodiments of the present invention are illustrated or described as block diagrams, flowcharts, or represented using certain other images, it will be understood that the blocks, apparatuses, systems, techniques, or methods described herein can be implemented as non-limiting examples in hardware, software, firmware, dedicated circuitry or logic, general-purpose hardware or controllers or other computing devices, or certain combinations thereof.

[0035] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0036] While the specific embodiments of the present invention have been described above, they are not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art without creative effort based on the technical solutions of the present invention are still within the scope of protection of the present invention.

Claims

1. A fault identification method based on smart measurement switch, characterized in that, The method includes the following steps: Step S1: Acquire multi-source data, preprocess the multi-source data, and output a standardized feature dataset; the multi-source data includes electrical data collected by the smart metering switch and meter data received from the connected smart meters. Step S2: Using the standardized feature dataset as input, the edge computing model is triggered by anomaly feature matching to output preliminary fault results, which include fault type, fault location, and confidence level. The step of triggering the edge computing model through anomaly feature matching specifically includes: based on the standardized feature dataset, when an anomaly feature matching any edge computing model is detected, the matching edge computing model outputs the preliminary fault result; The edge computing model includes at least a transient mutation calculation model, a time-series correlation analysis model, and a power quality anomaly identification model. The transient fault calculation model identifies transient faults based on transient characteristic parameters, which include at least voltage surges, voltage drops, transient duration, and instantaneous rate of change of voltage or current. The identification of transient faults specifically includes: comparing the transient characteristic parameters with a joint threshold of transient characteristics to match the electrical characteristics of the transient fault and obtain a predicted transient fault; simultaneously acquiring the protection action of the intelligent measurement switch, cross-validating the protection action with the predicted transient fault, eliminating false faults, and outputting a preliminary result of the transient fault containing the fault type and transient characteristic parameters. The time-series correlation analysis model identifies power outage faults based on power outage feature parameters, which include at least the power outage timestamp of the meter, the characteristics of the voltage loss event cluster, and the voltage status of the switch bus. The power quality anomaly identification model identifies power quality faults based on power quality characteristic parameters, which include at least three-phase voltage, phase sequence characteristics, and neutral current. The power quality anomaly identification model calculates phase data and three-phase imbalance based on the power quality characteristic parameters to identify phase loss and phase failure power quality faults. The edge computing model also includes a topology linkage correction model, which acquires topology identification signals at a preset period, generates dynamic topology relationships and updates the transformer area topology map, and maintains meter files based on the transformer area topology map. The meter files include at least the meter number, meter location and communication link status. Step S3: Obtain topology identification signals from the smart meter connected below according to a preset cycle, and obtain the dynamic topology relationship between the smart meter and the smart meter based on the topology identification signals; Step S4: Verify the dynamic topology relationship and the preliminary fault results to generate a fault assessment result; Step S4 specifically includes: Obtain preliminary fault results of the edge computing model, perform spatial verification based on the dynamic topology relationship, compare the preliminary fault results that pass the verification, and output the fault judgment result with the highest confidence. The fault judgment result includes at least fault type, fault location and confidence. 2.The smart metering breaker based fault identification method of claim 1, wherein, The preprocessing described in step S1 includes: The multi-source data is aligned to form a time-series synchronized dataset. The timing synchronization dataset is processed using a sliding window filtering algorithm to remove instantaneous interference values. For different types of faults, corresponding fault feature parameters are extracted, and the fault feature parameters are normalized to output a standardized feature dataset.

3. A fault identification system based on smart metering switches, characterized by, To implement the fault identification method as described in any one of claims 1-2, the system is deployed on an intelligent measuring switch, and the system includes a multi-source data acquisition module, a data preprocessing module, and an edge computing module; wherein: The multi-source data acquisition module is used to acquire electrical data through high-frequency electrical signals, and to receive meter data from the connected smart meters. The data preprocessing module is used to preprocess the multi-source data and output a standardized feature dataset; The edge computing module includes a fault identification unit, a topology identification unit, and a fusion judgment unit. The fault identification unit is used to take the standardized feature dataset as input, trigger the edge computing model through abnormal feature matching, and output preliminary fault results. The topology identification unit has a built-in topology identification module, which is used to obtain topology identification signals from the smart meter below according to a preset period, and obtain the dynamic topology relationship between the smart meter and the smart meter based on the topology identification signals. The fusion judgment unit is used to verify the dynamic topology relationship and the preliminary fault results, and generate fault judgment results.

4. A computer-readable storage medium, characterized in that, The computer-readable storage medium has a computer program that, when executed by a processor, implements the fault identification method based on an intelligent measuring switch as described in any one of claims 1-2.

Images