Fault identification system of wide-area monitoring network

Abstract

The invention relates to a fault identification system for a wide-area monitoring network. The system comprises a load side acquisition device, an edge computing terminal, a cloud analysis platform and a wireless communication module, the load side acquisition device is deployed at a neutral point of a secondary side of a distribution transformer or a user wire inlet end, and is used for acquiring amplitude, phase and frequency information of three-phase voltage in real time and transmitting the acquired data to the edge computing terminal through a wireless communication module; the plurality of load side acquisition devices cover each key node of the power distribution network to form a wide-area monitoring network; the edge computing terminal preprocesses the collected data and sends a preprocessing result to the cloud analysis platform; and the cloud analysis platform receives the preprocessing result of the edge computing terminal, and completes global fault identification by using a fusion model of a support vector machine and a deep neural network. Compared with the prior art, the method has the advantages of realizing rapid fault identification of the wide-area monitoring network and the like.

Description

Technical Field

[0001] This invention relates to the field of power distribution network fault identification technology, and in particular to a fault identification system for a wide-area monitoring network. Background Technology

[0002] As the end point of the power system, the distribution network directly faces users, and its power supply reliability and security are directly related to people's well-being and socio-economic operation. However, distribution network lines are complex in topology and have variable environments, resulting in frequent faults such as line breaks and grounding. In particular, compound faults involving single-phase line breaks and grounding not only cause power outages for users but also create dangerous voltages due to exposed or fallen conductors, becoming a major cause of electric shock accidents to people and animals. With the acceleration of urbanization and the improvement of public safety awareness, distribution network faults have evolved from simple technical issues into sensitive events involving public safety and social stability. Against this background, the core purpose of researching high-precision line break fault identification technology is to extract transient and steady-state characteristics of faults (such as voltage phase shift and traveling wave propagation characteristics) to construct differentiated fault type discrimination models and quickly distinguish between line break faults, grounding faults, and compound faults. This technology aims to overcome the limitations of traditional single criteria, solve the problems of misjudgment and missed judgment caused by similar fault characteristics, and thus accurately locate the fault point and identify the risk type. This provides a basis for decision-making for rapid fault isolation and power restoration. At the same time, by providing real-time early warning of high-risk faults (such as grounding failure), it minimizes the risk of personal injury and promotes the upgrade of power distribution network operation and maintenance from "passive response" to "active defense", thus consolidating the safety bottom line of power supply.

[0003] In-depth research into distribution network fault identification technology holds multiple strategic significances for the sustainable development of the power industry and public safety. From a technical perspective, the approach integrating wide-area synchronous measurement, artificial intelligence, and high-precision signal processing can not only achieve accurate fault type identification but also accumulate massive amounts of fault waveform data. This provides data support for distribution network condition assessment, equipment lifespan prediction, and the development of new protection devices, contributing to the construction of a "observable, measurable, and controllable" smart distribution network system. In the context of energy transition, with the high penetration rate of distributed photovoltaic power and electric vehicle charging piles, distribution network fault characteristics are becoming increasingly complex, making traditional identification methods inadequate for the challenges of multiple power sources and scenarios. Developing new technologies can effectively address the interference caused by fluctuations in new energy sources, ensuring stable grid operation and promoting the consumption of clean energy. From a social benefit perspective, rapid and accurate fault identification can significantly shorten power outage time, reduce economic losses for critical users such as hospitals and transportation hubs, and enhance urban operational resilience. Simultaneously, by avoiding the persistence of high-risk faults such as grounding failures, it significantly reduces the probability of electric shock accidents, safeguarding public life and property and demonstrating the social responsibility of power companies. Furthermore, in terms of public opinion management, the transparent application of technology can enhance public trust in electricity services, avoid social criticism caused by delays in fault handling, maintain corporate brand image, and inject technological momentum into building a harmonious power supply and consumption relationship. Currently, the power distribution network lacks a low-cost, closed-loop fault identification system.

[0004] A search revealed Chinese invention patent application publication number CN117454234A, which discloses a method and device for fault identification in county-level power grids based on cloud-edge collaboration. The device includes: a data acquisition module for acquiring real-time data from terminal equipment and distribution network lines and inputting it into an edge computing node; a fault identification preprocessing module for the edge computing node to perform fault identification preprocessing based on an existing fault identification model and transmit the results to a cloud platform; and an inference model building module for the cloud platform to combine terminal data and edge data to optimize the model, build an inference model, and enrich the edge identification system. This existing patent application suffers from problems due to the use of a power waveform recording system to acquire real-time data from terminal equipment and distribution network lines, resulting in asynchronous acquisition, limited sampling accuracy and speed, and consequently, inaccurate fault identification.

[0005] How to achieve rapid fault identification and closed-loop processing in the power distribution network has become a technical problem that needs to be solved. Summary of the Invention

[0006] The purpose of this invention is to overcome the shortcomings of the existing technology and provide a fault identification system for a wide-area monitoring network.

[0007] The objective of this invention can be achieved through the following technical solutions: According to one aspect of the present invention, a fault identification system for a wide-area monitoring network is provided, the system comprising a plurality of load-side acquisition devices, multiple edge computing terminals, a cloud analysis platform and a wireless communication module; Load-side acquisition devices are deployed at the neutral point of the secondary side of the distribution transformer or at the user's incoming line. They are used to collect current and voltage signals in the distribution network in real time and transmit the collected data to the edge computing terminal through a wireless communication module. Several load-side acquisition devices cover each key node of the distribution network, forming a wide-area monitoring network. The edge computing terminal is responsible for preprocessing the data collected by the load-side acquisition device, including feature extraction and preliminary classification, and sending the preprocessed results to the cloud analysis platform for further processing. The cloud-based analytics platform receives preprocessed results from edge computing terminals and uses a fusion model of support vector machines and deep neural networks to complete global fault identification.

[0008] Preferably, the load-side acquisition device includes: The inductive power supply powers the entire load-side data acquisition device. The sensor module is responsible for accurately measuring the current and voltage signals in the distribution network lines, including a current sensing unit and a voltage sensing unit. The data acquisition and processing module integrates high-speed sampling and storage functions, and can automatically sample the output data of the sensor module and store the transient traveling waves of power distribution network faults. The voltage sensing unit is used to collect the amplitude, phase, and frequency information of the three-phase voltage in real time. A current sensing unit is used to collect the current of the power distribution line in real time. It is a communication module responsible for data interaction between load-side acquisition devices, and between the load-side acquisition devices and the monitoring backend and edge computing terminal. More preferably, the current sensing unit includes a printed circuit board Rogowski coil and a signal conditioning and amplification circuit. The power distribution line passes through the printed circuit board Rogowski coil, sensing the line current and generating a current signal, which is then conditioned and sampled.

[0009] More preferably, the voltage sensing unit includes a capacitive voltage divider sensor, which adopts the distributed voltage measurement principle, monitors the voltage on the sampling capacitor through a differential amplifier, and samples the voltage signal after signal conditioning.

[0010] More preferably, the magnetic core of the inductive power supply obtains energy from the current in the power distribution line through electromagnetic induction, and the induced output power is positively correlated with the primary current.

[0011] More preferably, when the primary current is higher than the starting current corresponding to the power consumed by the load-side acquisition device, all the power consumed by the load-side acquisition device is provided by the inductive output power, and the additional power obtained by induction is stored in the backup lithium battery; when the primary current is lower than the starting current corresponding to the power consumed by the load-side acquisition device, the power is switched to the backup lithium battery.

[0012] Preferably, the preprocessing further includes performing sequence component analysis based on voltage data to obtain the sequence component phase difference, wavelet packet decomposition to calculate the frequency domain energy distribution, and real-time calculation of the phase difference change rate between the faulty phase and the non-faulty phase to identify the slow-changing characteristics of the open circuit fault.

[0013] More preferably, the time-domain statistics, frequency-domain energy distribution, and sequence component phase difference of the load-side voltage are input into the fusion model to complete global fault type identification and risk assessment.

[0014] Preferably, the fusion model is dynamically updated through an online learning mechanism to adapt to changes in power grid topology and fluctuations in new energy sources.

[0015] Preferably, the system further includes: Early warning and execution unit: responsible for converting fault identification results into specific operation and maintenance actions, including pushing fault information to relevant personnel, and also supporting linkage with the power distribution automation system to automatically trigger circuit breakers to perform fault section isolation operations; GPS synchronization module: Used to ensure that all monitoring points in the wide-area monitoring network are on the same time reference.

[0016] Compared with the prior art, the present invention has the following beneficial effects: 1) In this invention, the load-side acquisition device is deployed at the neutral point of the secondary side of the distribution transformer or at the user's incoming line, covering all key nodes of the distribution network. It collects voltage and current data of the wide-area monitoring network in real time. After the collected data is preprocessed by the edge computing terminal, it is then analyzed by the cloud platform using a fusion model of support vector machine and deep neural network to complete global fault identification. Compared with traditional wide-area measurement that relies on high-precision GPS synchronization and high-frequency transient data, the load-side acquisition device deployed locally in this invention collects data in real time and only needs power frequency cycle data to achieve rapid identification, reducing the need for high-precision synchronization and massive data. This low data dependence feature greatly reduces the need for high-speed communication network and big data storage, making it more suitable for the distributed architecture of the distribution network with "many points and wide coverage".

[0017] 2) This invention employs advanced high-energy-density power extraction technology, which can efficiently extract energy from distribution network lines and provide stable power support for the acquisition device. Compared with traditional power extraction methods, this technology can still ensure sufficient energy supply under low current loads, ensuring that the device can work normally under various operating conditions and will not experience measurement interruptions or data loss due to insufficient energy, thereby improving the reliability and continuity of fault transient waveform capture; 3) The load-side voltage and current monitoring of this invention is less affected by interference from renewable energy grid connection. By analyzing the voltage characteristics downstream of the renewable energy grid connection point, the fault type can be effectively identified, avoiding misjudgment problems caused by renewable energy disconnection or changes in control strategies, and improving the reliability of fault handling in complex power scenarios with a high proportion of renewable energy access.

[0018] 4) The fault identification system of the present invention can improve the safety of operation and maintenance and social benefits: Traditional fault identification methods rely on manual inspection after a fault occurs, which poses a risk to personal safety; while the present invention, based on real-time monitoring of load side voltage and current, can provide early warning of high-risk faults such as wire breakage and grounding, and guide operation and maintenance personnel to accurately locate risk points. This not only shortens the fault investigation time, but also reduces the probability of electric shock accidents through proactive prevention and control. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the fault identification system in this invention; Figure 2 This is a schematic diagram of the installation of the load-side acquisition device in this invention; Figure 3 This is a schematic diagram of the load-side acquisition device in this invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0021] Example 1 This embodiment relates to a fault identification system for a wide-area monitoring network. It aims to accurately identify single-phase open circuits, high-resistance grounding, and compound faults by real-time monitoring of the three-phase voltage signals on the distribution transformer side or user end, combined with advanced signal processing and artificial intelligence algorithms. It also generates risk level assessment and location information simultaneously, forming a closed-loop fault handling process from data acquisition and feature analysis to intelligent decision-making.

[0022] The system adopts a distributed architecture design, such as Figure 1The system comprises a load-side voltage and current acquisition module, several edge computing terminals, a cloud analysis platform, a wireless communication module, an early warning and execution unit, and a GPS synchronization module, forming a complete closed loop from data acquisition, feature extraction, intelligent classification to fault response. Compared to traditional power source-side current analysis methods, this system offers higher anti-interference capabilities, lower hardware costs, and stronger scenario adaptability, making it particularly suitable for complex distribution network scenarios with a high proportion of renewable energy integration. Through modular design and edge-cloud collaborative computing, the system achieves a balance between real-time performance, accuracy, and economy in fault identification, providing reliable technical support for the intelligent transformation of distribution networks.

[0023] Load-side data acquisition device: such as Figure 2 The load-side data acquisition device is the data source of this system, deployed at the neutral point of the secondary side of the distribution transformer or at the user's incoming line. It is used to accurately capture and analyze transient waveforms during distribution network faults. The load-side data acquisition device consists of five main modules: an inductive power supply, a sensor module, a data acquisition and processing module, a communication module, and a power management module. These five modules work together, each with a clearly defined function, to complete the entire process from energy acquisition and signal processing to data transmission. The load-side data acquisition device has a built-in signal conditioning circuit that filters, amplifies, and performs AD conversion on the raw voltage signal, with a sampling rate of 10 kHz, meeting the requirements for power frequency and transient characteristic analysis. Furthermore, the load-side data acquisition device supports local storage, recording complete waveform data from 5 cycles before the fault to 20 cycles after the fault, providing a high-quality data foundation for subsequent feature extraction and fault analysis. Through distributed deployment, the load-side data acquisition device can cover all key nodes of the distribution network, forming a wide-area monitoring network and providing comprehensive data support for fault identification.

[0024] Edge computing terminal: The edge computing terminal is the core processing unit of the system, responsible for preprocessing the collected data, namely, real-time feature extraction and preliminary classification. The terminal integrates a low-power ARM Cortex-M7 processor, signal conditioning circuitry, an AD conversion module, and local storage, possessing powerful data processing capabilities and low power consumption. The edge computing terminal adopts a modular design, supporting various signal processing algorithms, including sequence component analysis, wavelet packet decomposition, and phase difference dynamic tracking. For example, by calculating positive-sequence, negative-sequence, and zero-sequence voltage components, the proportion of negative-sequence voltage is extracted (e.g., the proportion of negative-sequence voltage exceeds 15% in a single-phase open circuit); a three-level decomposition using the DB4 wavelet basis is performed to calculate the energy entropy value (i.e., wavelet packet entropy) of each frequency band, characterizing the transient features of the fault; and the phase difference change rate between the faulty phase and the non-faulty phase is calculated in real time to identify the slowly changing characteristics of the open circuit fault. The local processing capability of the edge computing terminal significantly reduces data transmission volume, lowers communication costs, and simultaneously improves the real-time performance and reliability of the system.

[0025] Cloud-based analytics platform: The cloud-based analytics platform serves as the system's intelligent decision-making center. It receives preprocessed data from multi-node edge computing terminals and employs a fusion model of Support Vector Machine (SVM) and Deep Neural Network (DNN). Input features include time-domain statistics of load-side voltage, frequency-domain energy distribution, and sequence component phase difference to complete global fault type identification and risk assessment. The fusion model is dynamically updated through an online learning mechanism, adapting to changes in grid topology and fluctuations in new energy sources, achieving a classification accuracy significantly higher than traditional threshold-based methods. The cloud-based analytics platform utilizes a distributed computing architecture, supporting parallel processing and storage of massive amounts of data, ensuring stable operation under high-concurrency scenarios. SVM handles linear classification based on sequence components and phase differences (e.g., distinguishing between grounding and open-circuit faults), while DNN processes high-dimensional nonlinear features such as wavelet packet entropy (e.g., identifying composite faults). To address the imbalanced fault sample problem, the platform introduces an adaptive weighted loss function and Synthetic Minority Oversampling Technology (SMOTE) to improve the sensitivity for identifying rare fault types (e.g., open-circuit grounding). In addition, the cloud-based analytics platform supports online model updates, continuously optimizing parameters through incremental learning to adapt to changes in power grid topology and fluctuations in new energy sources, ensuring a continuous improvement in classification accuracy.

[0026] Wireless Communication Module: The communication module serves as the bridge for system data transmission, employing dual-mode communication technologies of LoRa and NB-IoT to support data interaction between local terminals (edge ​​computing terminals) and the cloud (cloud analytics platform). LoRa is suitable for long-distance, low-power rural and remote areas, while NB-IoT provides highly reliable urban coverage, ensuring stable system operation under various network conditions. The communication module design considers data security and transmission efficiency, employing encryption protocols and compression algorithms to ensure the integrity and confidentiality of critical data during transmission. Furthermore, the module supports breakpoint resumption and data caching functions, guaranteeing data integrity and continuity even during network interruptions. Through optimized communication protocols and network configuration, the system achieves low-latency, highly reliable data transmission, providing robust communication support for fault identification.

[0027] Early warning and execution unit: Early warning and execution unit (not in Figure 1The early warning and execution unit (as shown in the diagram) is the final output stage of the system, responsible for translating fault identification results into specific maintenance actions. It pushes fault type (e.g., "C-phase open circuit with grounding"), location (based on topology correlation analysis), and risk level (e.g., high-risk, medium-risk, low-risk) information via SMS, SCADA system interface, or mobile app to guide maintenance personnel in quickly locating and handling faults. Simultaneously, the early warning and execution unit supports linkage with the distribution automation system, automatically triggering circuit breakers to perform fault section isolation operations, reducing manual intervention and improving fault handling efficiency. Early warning information uses a tiered push mechanism, prioritizing high-risk faults (e.g., open circuit with grounding) to ensure timely handling of critical faults. Furthermore, the unit has a built-in user feedback function, allowing maintenance personnel to submit fault handling results through the interface. The system automatically updates the fault identification knowledge base, optimizes model performance, and forms a closed-loop optimization mechanism.

[0028] GPS Synchronization Module: The GPS synchronization module is a key technological component ensuring that all monitoring points maintain the same time reference. By integrating a high-precision GPS receiver, this module provides the system with microsecond-level time synchronization accuracy, which is crucial for accurately measuring and comparing voltage and current waveforms at different locations. The GPS synchronization module not only provides timestamps for data acquisition but also corrects time errors at each monitoring point through synchronization signals, ensuring data consistency and accuracy. Furthermore, this module supports the BeiDou Navigation Satellite System (BDS) as a backup time source, improving system reliability and redundancy. In practical applications, the GPS synchronization module, through close collaboration with the main control CPU module, achieves precise time synchronization across the entire network, providing a solid time reference for subsequent fault location and analysis.

[0029] By employing load-side voltage characteristic analysis, edge-cloud collaborative computing, and intelligent classification algorithms, this system achieves high-precision identification and rapid response to distribution network faults. Adopting a modular design, it is compatible with existing smart meters and distribution terminal equipment, and features low cost, easy deployment, and wide adaptability. Compared to traditional methods, this system significantly improves anti-interference capabilities, classification accuracy, and scenario adaptability, making it particularly suitable for complex distribution network scenarios with a high proportion of renewable energy integration. Through real-time early warning and automatic isolation, the system substantially reduces the risk of electric shock accidents and fault recovery time, providing strong technical support for the safe operation and intelligent transformation of distribution networks.

[0030] Example 2 This embodiment relates to a fault identification system for a wide-area monitoring network, the system framework of which is as follows: Figure 1As shown, several load-side data acquisition devices are distributed along the power distribution line. Each device uses a GPS clock as its local clock reference to ensure synchronized data acquisition at each monitoring point. The load-side data acquisition devices can acquire voltage and current signals on the line in real time and transmit the acquired waveform data to the edge computing terminal via wireless networks such as GPRS / CDMA / 4G / 5G through various methods including timed triggering, call triggering, overcurrent, overvoltage, and abnormal event triggering. In terms of sensor design, current monitoring uses a PCB Rogowski coil as the current sensor, while voltage measurement is achieved through a voltage sensing unit utilizing wires and distributed capacitance, such as... Figure 2 As shown, the edge computing terminal determines the line section where the fault is located by matching and analyzing the fault components uploaded from different monitoring points with the line characteristics. It then further analyzes the fault point on the cloud platform based on the traveling wave of the fault current, and finally completes the fault type identification by combining the transient voltage and current characteristics of the fault.

[0031] like Figure 3 The load-side data acquisition device includes an inductive power supply, a sensor module, a data acquisition and processing module, a communication module, and a power management module. The following is a detailed description of each module.

[0032] Sensor Module: The sensor module is responsible for accurately measuring the current and voltage signals in the distribution network lines, including current sensing units and voltage sensing units. The current sensing unit contains a printed circuit board Rogowski coil and signal conditioning amplifier circuit. The distribution line passes through the Rogowski coil, which can sensitively sense the line current and generate a corresponding current signal. After conditioning, the signal is sent to the data acquisition and processing module for sampling. Its measurement range is 800A. It has advantages such as fast response speed, high measurement accuracy, and wide bandwidth, and can effectively capture rapidly changing current waveforms and high-frequency components in transient processes. The voltage sensing unit is based on the distributed voltage measurement principle. It detects the voltage on the sampling capacitor through a differential amplifier, and after filtering, amplification, and level conversion, the signal is sent to the data acquisition and processing module for sampling. It can accurately extract conductor voltage signals in complex electromagnetic environments, unaffected by environmental changes, providing accurate voltage data for fault location and identification.

[0033] Data Acquisition and Processing Module: This module is the intelligent hub of the device, using a low-power MSP430F5438 processor as the main controller. It integrates high-speed sampling and storage functions, automatically initiating and storing transient current traveling waves. Working in conjunction with a GPS synchronization module, it ensures synchronous measurement, guaranteeing data acquisition from all detection points at the same time reference, providing time assurance for subsequent wide-area measurement and analysis. To ensure no critical data is lost, the module is equipped with a ferroelectric memory for storing recorded current waveform data. This module boasts extremely high sampling capabilities, with a power frequency current sampling rate of 12.8 kHz, a traveling wave sampling rate of no less than 1 MHz, and a sampling accuracy of 12 bits, accurately capturing every detail of the fault.

[0034] The inductive power supply powers the entire load-side data acquisition device. Its magnetic core extracts energy from the power distribution line current through electromagnetic induction, and its induced output power is positively correlated with the primary current. When the power distribution line current (primary current) is higher than the starting current corresponding to the power consumed by the load-side data acquisition device (the minimum primary current required for the inductive power supply to achieve the minimum load power), all the power consumed by the load-side data acquisition device is provided by the inductive output power. The additional power obtained through induction is then stored in the backup lithium battery via PWM charging under the control of the power management module. When the line current is lower than the starting current, the power supply switches to the backup lithium battery. The inductive power supply has a pre-amplifier power control and protection module to protect the downstream circuitry from overvoltage damage under high current conditions. The rectifier and filter management module converts the AC power induced by the magnetic core into DC power. The inductive power supply adopts a two-stage DC / DC design for convenient power management, with a final output voltage of 3.3V and a maximum output power of 2.5W. The power management module uses the C8051F320 as the main control CPU. It takes the voltage at key locations in the power circuit as the input signal and outputs switching control signals according to the management strategy to rationally allocate the acquired energy. The C8051F320 interacts with the MSP430F5438 data acquisition and processing module via asynchronous serial communication to complete the communication tasks for each phase as defined by a custom communication protocol.

[0035] Communication Module: The communication module is responsible for data interaction between load-side acquisition devices, between load-side acquisition devices and the monitoring backend, and between the communication module and the edge computing terminal. The communication module consists of two parts: a Zigbee short-range wireless communication unit and a GPRS / CDMA / 4G / 5G wireless communication unit. The Zigbee module is used to complete data interaction between the various load-side acquisition devices on the three-phase line, ensuring the synchronization and integrity of the three-phase data. The wireless communication unit acts as a long-range communication unit, responsible for transmitting the acquired waveform files and monitoring data to the edge computing terminal via a wireless network. The device supports multiple triggering methods such as timing, call, overvoltage, overcurrent, and anomaly, enabling flexible data reporting.

[0036] Power Management Module: The power management module is responsible for the comprehensive management and allocation of power to the load-side data acquisition devices. This module can monitor the power supply voltage in real time and rationally allocate energy according to the device's operating status and needs, ensuring that each module of the load-side data acquisition devices receives a stable and reliable power supply under different operating conditions. Simultaneously, the power management module also has a control signal output function, communicating with the EFM32 Giant Gecko processor via a UART interface to achieve precise control and adjustment of the operating status of the load-side data acquisition devices. Through the coordinated work of the above five modules, this invention can achieve accurate capture and analysis of transient waveforms of distribution network faults, significantly improving fault location accuracy and speed, enhancing system reliability and stability, optimizing operation and maintenance efficiency and decision support, and providing strong technical support for the safe and reliable operation of the distribution network.

[0037] Through the coordinated operation of the five modules mentioned above, a series of problems existing in traditional distribution networks are solved. These problems include insufficient energy supply, low measurement accuracy, data asynchrony, poor communication reliability, and poor system stability, which are inherent in traditional distribution network fault transient waveform capture technologies. This provides a comprehensive technical solution for the safe and reliable operation of distribution networks. Furthermore, the system supports expansion and dynamic optimization, making it particularly suitable for diagnosing mechanical and electrical faults under complex operating conditions, providing strong technical support for the safe and reliable operation of modern distribution networks. This highly integrated design not only gives the entire system powerful data processing capabilities and stability but also facilitates expansion and maintenance, laying a solid foundation for future intelligent upgrades.

[0038] The load-side data acquisition device supports low-cost and large-scale deployment. Compared to existing technologies that require the installation of traveling wave sensors or high-precision synchronous measurement units, load-side voltage and current monitoring can be achieved by upgrading existing secondary side measurement devices of distribution transformers or smart meters on the user side, without the need for additional high-cost hardware. This is particularly suitable for the renovation of old distribution networks and applications in remote areas. The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in the present invention, and these modifications or substitutions should all be covered within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A fault identification system for a wide-area monitoring network, characterized in that, The system includes several load-side data acquisition devices, multiple edge computing terminals, a cloud analytics platform, and a wireless communication module; The load-side acquisition device is deployed at the neutral point of the secondary side of the distribution transformer or at the user's incoming line. It is used to collect current and voltage signals in the distribution network line in real time and transmit the collected data to the edge computing terminal through a wireless communication module. Several load-side acquisition devices cover key nodes of the power distribution network, forming a wide-area monitoring network; The edge computing terminal is responsible for preprocessing the data collected by the load-side acquisition device, including feature extraction and preliminary classification, and sending the preprocessed results to the cloud analysis platform for further processing. The cloud-based analytics platform receives preprocessed results from edge computing terminals and uses a fusion model of support vector machines and deep neural networks to complete global fault identification.

2. The fault identification system for a wide-area monitoring network according to claim 1, characterized in that, The load-side acquisition device includes: The inductive power supply powers the entire load-side data acquisition device. The sensor module is responsible for accurately measuring the current and voltage signals in the distribution network lines, including a current sensing unit and a voltage sensing unit. The data acquisition and processing module integrates high-speed sampling and storage functions, and can automatically sample the output data of the sensor module and store the transient traveling waves of power distribution network faults. The communication module is responsible for data interaction between load-side acquisition devices, and between load-side acquisition devices and the monitoring backend and edge computing terminals.

3. The fault identification system for a wide-area monitoring network according to claim 2, characterized in that, The current sensing unit includes a printed circuit board Rogowski coil and a signal conditioning and amplification circuit. The power distribution line passes through the printed circuit board Rogowski coil, senses the line current and generates a current signal, which is then conditioned and sampled.

4. The fault identification system for a wide-area monitoring network according to claim 2, characterized in that, The voltage sensing unit includes a capacitive voltage divider sensor, which adopts the distributed voltage measurement principle. It monitors the voltage on the sampling capacitor through a differential amplifier and samples the voltage signal after signal conditioning.

5. The fault identification system for a wide-area monitoring network according to claim 2, characterized in that, The magnetic core of the inductive power supply obtains energy from the current in the power distribution line through electromagnetic induction, and the induced output power is positively correlated with the primary current.

6. The fault identification system for a wide-area monitoring network according to claim 5, characterized in that, When the primary current is higher than the starting current corresponding to the power consumed by the load-side acquisition device, all the power consumed by the load-side acquisition device is provided by the inductive output power, and the extra power obtained by induction is stored in the backup lithium battery; when the primary current is lower than the starting current corresponding to the power consumed by the load-side acquisition device, the power supply is switched to the backup lithium battery.

7. The fault identification system for a wide-area monitoring network according to claim 1, characterized in that, The preprocessing also includes performing sequence component analysis based on voltage data to obtain the sequence component phase difference, wavelet packet decomposition to calculate the frequency domain energy distribution, and real-time calculation of the phase difference change rate between the faulty phase and the non-faulty phase to identify the slow-changing characteristics of the open circuit fault.

8. The fault identification system for a wide-area monitoring network according to claim 7, characterized in that, The time-domain statistics, frequency-domain energy distribution, and sequence component phase difference of the load-side voltage are input into the fusion model to complete global fault type identification and risk assessment.

9. The fault identification system for a wide-area monitoring network according to claim 1, characterized in that, The fusion model is dynamically updated through an online learning mechanism to adapt to changes in power grid topology and fluctuations in new energy sources.

10. The fault identification system for a wide-area monitoring network according to claim 1, characterized in that, The system also includes: Early warning and execution unit: responsible for converting fault identification results into specific operation and maintenance actions, including pushing fault information to relevant personnel, and also supporting linkage with the power distribution automation system to automatically trigger circuit breakers to perform fault section isolation operations; GPS synchronization module: Used to ensure that all monitoring points in the wide-area monitoring network are on the same time reference.

Images