CN119667370A

CN119667370A - Real-time fault detection and automatic repair system for power grid

Info

Publication number: CN119667370A
Application number: CN202411753608.1A
Authority: CN
Inventors: 文敬德; 叶剑锋; 龚艳玲; 李扬; 龚廷; 彭军; 张岩鹏; 苏高扬; 陈俊学; 杜克
Original assignee: Yichang Power Supply Co of State Grid Hubei Electric Power Co Ltd; Xiangyang Power Supply Co of State Grid Hubei Electric Power Co Ltd
Current assignee: Yichang Power Supply Co of State Grid Hubei Electric Power Co Ltd; Xiangyang Power Supply Co of State Grid Hubei Electric Power Co Ltd
Priority date: 2024-12-02
Filing date: 2024-12-02
Publication date: 2025-03-21

Abstract

The invention relates to a real-time fault detection and automatic repair system for a power grid, which comprises a multi-level sensing module, a data processing module, an automatic repair module and a monitoring and log module. The multi-level sensing module is used for acquiring data such as voltage, current, frequency and temperature in the running process of the power grid by deploying multi-type sensors and sending the data to the data processing module. The data processing module analyzes the acquired data through an embedded intelligent algorithm, identifies potential faults and generates fault diagnosis results. And the automatic repair module executes repair operations of isolating a fault region, reconstructing a power grid topological structure and recovering normal power supply according to the fault diagnosis result. The monitoring and logging module records the running state of the power grid, fault diagnosis and repair processes, and provides fault analysis and performance evaluation functions through a user interface. The system realizes real-time detection, accurate positioning, automatic repair and comprehensive monitoring of operation data of the power grid fault, and obviously improves the safety and reliability of the operation of the power grid.

Description

Real-time fault detection and automatic repair system for power grid

Technical Field

The invention relates to the technical field of electric power, in particular to a real-time fault detection and automatic repair system for a power grid.

Background

In the prior art, various faults may occur in the operation of the power grid, including short circuits, overload, equipment anomalies, and the like. In order to improve the reliability of the power grid, the existing system collects power grid operation data through sensors and analyzes the state of the power grid operation data, and part of the system can locate faults and process the faults in a manual or semi-automatic mode. In addition, some intelligent systems can provide basic monitoring functions of the operation of the power grid, but are often limited to data recording after failure occurrence and processing of single failures.

However, the prior art still has obvious problems, which are mainly reflected in insufficient real-time performance of fault detection, insufficient accuracy of diagnosis results and low automation degree of the repair process. The existing system has limitation on the identification of fault types and the evaluation of influence ranges, meanwhile, the topology reconstruction and repair operation of the power grid often need manual intervention, and the requirements of the modern power grid on quick recovery and high reliability are difficult to meet.

In order to solve the problems, the invention provides a novel power grid real-time fault detection and automatic repair system.

Disclosure of Invention

The application provides a real-time fault detection and automatic repair system for a power grid, which is used for improving the safety and reliability of the operation of the power grid.

The application provides a real-time fault detection and automatic repair system for a power grid, which comprises the following components:

the multi-level sensing module is used for collecting operation data of the power grid through the multi-type sensors and sending the collected operation data to the data processing module, wherein the operation data comprises voltage, current, frequency and temperature;

The data processing module is used for receiving the operation data provided by the multi-level perception module, analyzing the operation data through an embedded intelligent algorithm to identify potential faults, and generating a fault diagnosis result when the faults are detected, wherein the fault diagnosis result comprises a fault type, a fault position and an influence range;

the automatic repair module is used for executing repair operation according to the fault diagnosis result generated by the data processing module, wherein the repair operation comprises fault region isolation, power grid topological structure reconstruction and normal power supply restoration;

The monitoring and logging module is used for recording the running state of the power grid, fault diagnosis results and repair operation processes, generating a historical data report and providing fault analysis and running performance evaluation through a user interface.

Further, the multi-level sensing module is specifically configured to:

Respectively deploying a voltage sensor and a current sensor at a transformer substation, a power transmission and distribution line and a load end, and collecting voltage and current data of each node in real time, wherein the voltage sensor is used for detecting an instantaneous value of the voltage of the node, and the current sensor is used for monitoring the amplitude and the direction of current;

Installing temperature sensors at the positions of high-load equipment and key circuits, detecting the temperature rise conditions of equipment shells and circuits at fixed time, and sending temperature data to a data processing unit through a wireless communication module;

The method comprises the steps of continuously sampling the change of system frequency at the integrated frequency acquisition device of the main transformer substation, capturing frequency offset caused by load fluctuation or faults, packaging all acquired data and transmitting the packaged data to the data processing module.

Still further, the automated repair module is specifically configured to:

starting a circuit breaker control unit based on a fault diagnosis result, isolating the identified fault region, wherein the isolating comprises physically disconnecting a fault line or device from a normal power supply portion;

Calculating a power grid topological structure of a non-fault area, generating a new power supply path by starting a standby power supply, a standby line or redistributing an existing load, and sending a command to related switch equipment through a communication module to execute topology reconstruction;

after the topology adjustment is confirmed, the voltage class and the power supply power are adjusted through the transformer, and the power supply of the normal load area is gradually restored.

Further, the monitoring and logging module is specifically configured to:

acquiring real-time operation data comprising voltage, current, frequency and temperature parameters through a state acquisition unit, recording abnormal data when a fault occurs, and generating operation state snapshots before and after the fault;

in the fault repairing process, specific steps and execution time of isolation operation, topology adjustment and power supply recovery are recorded in real time;

After the repair is completed, a complete log report comprising power grid running state data, fault cause analysis and repair processes is generated, and is displayed and stored for an operator through a visual interface.

Furthermore, the intelligent algorithm adopted by the data processing module is realized by the following steps:

The method comprises the steps of utilizing wavelet transformation to decompose operation data collected by a multi-level perception module, extracting a low-frequency component by removing high-frequency noise, carrying out sliding window time sequence analysis on the low-frequency component, extracting characteristic data reflecting the operation state of a power grid, and generating the characteristic data of the power grid, wherein the extracted characteristic data comprises the change rate of a current waveform, the unbalance degree of voltage and the frequency offset characteristic;

Inputting the characteristic data into a pre-trained support vector machine classifier, dynamically optimizing a classification boundary based on a historical sample and real-time load data, and generating a diagnosis result of a fault type;

according to the analysis result of the voltage or current variation source, and combining the operation characteristics of equipment and lines to determine the specific position of the fault occurrence and generate a fault positioning result comprising fault equipment or line identification;

Based on fault positioning results, combining the connection relation and the real-time load state of equipment in a power grid, analyzing the fault influence range by means of gradual propagation, gradually calculating the operation state changes of a transformer substation, a power transmission line and load equipment and affected areas thereof, and generating detailed information of the complete affected areas when the influence is weakened to a preset negligible range, wherein the detailed information comprises equipment identification, operation parameter change values and specific influence ranges.

Furthermore, the decomposing the operation data collected by the multi-level sensing module by wavelet transformation, extracting the low frequency component by removing the high frequency noise includes:

The wavelet basis functions ψ _j,k (t) and Φ _j,k (t) are defined according to the following formula (1) and formula (2):

wherein t represents sampling time, j represents a scale parameter, k represents a time shift parameter;

ψ (t) and φ (t) are the basis functions of the wavelet functions, as shown in the following formulas (3) and (4):

ψ(t)=exp(-αt²)cos(2πβt) (3)

φ(t)=exp(-αt²) (4)

wherein alpha is a frequency offset adjustment parameter, beta is a waveform adjustment parameter, t is sampling time;

The operation data x (t) is subjected to multi-scale decomposition according to the following formulas (5) and (6), generating a high-frequency component D _j (t) and a low-frequency component a _j (t):

the high frequency component threshold T is calculated according to the following formula (7):

Wherein eta is a scale factor set based on the statistical property of the operation data of the power grid, N is the number of sampling points;

Zeroing a high-frequency component of which the absolute value is D _j (T) is greater than T to remove noise, and obtaining a denoised high-frequency component;

And reconstructing and extracting the low-frequency component reflecting the running state of the power grid by carrying out inverse wavelet transformation on the denoised high-frequency component and the low-frequency component, and taking the low-frequency component and the low-frequency component as key characteristic input for describing the state of the power grid.

The application has the following beneficial technical effects:

(1) According to the application, through an intelligent algorithm embedded in the data processing module and combining with multidimensional operation data acquired by the multilevel sensing module, potential faults in power grid operation can be accurately identified, and a diagnosis result comprising fault types, fault positions and influence ranges is generated, so that a reliable basis is provided for quick repair measures. (2) The automatic repair module can automatically execute isolation operation of a fault area according to a diagnosis result, quickly reconstruct a power grid topological structure, recover normal power supply, remarkably reduce power failure time and manual intervention requirements, and improve operation reliability and maintenance efficiency of a power grid. (3) Through the monitoring and log module, the system can record the running state, fault diagnosis result and repair operation process of the power grid in detail, generate a historical data report, provide visual fault analysis and performance evaluation through a user interface and support running optimization and maintenance decision of the power grid. (4) The system realizes the full-flow real-time response from fault detection to repair and has the capability of rapidly processing sudden faults. Meanwhile, due to the modularized design, all functional units are independent and mutually cooperated, and flexible configuration and expansion are facilitated according to the power grid scale and application requirements.

Drawings

Fig. 1 is a schematic diagram of a real-time fault detection and automatic repair system for a power grid according to a first embodiment of the present application.

Detailed Description

In the following description, numerous specific details are set forth in order to provide a thorough understanding of the present application. The present application may be embodied in many other forms than those herein described, and those skilled in the art will readily appreciate that the present application may be similarly embodied without departing from the spirit or essential characteristics thereof, and therefore the present application is not limited to the specific embodiments disclosed below.

The first embodiment of the application provides a power grid real-time fault detection and automatic repair system. Referring to fig. 1, a schematic diagram of a first embodiment of the present application is shown. The following describes a power grid real-time fault detection and automatic repair system in detail according to a first embodiment of the present application with reference to fig. 1.

The power grid real-time fault detection and automatic repair system comprises a multi-level sensing module 101, a data processing module 102, an automatic repair module 103 and a monitoring and log module.

The multi-level sensing module 101 is configured to collect operation data of the power grid through multiple types of sensors and send the collected operation data to the data processing module, where the operation data includes voltage, current, frequency and temperature.

The multi-level sensing module 101 is intended to efficiently and real-time collect key parameters of the operation of the power grid, including voltage, current, frequency and temperature, through various types of sensors. The module can be deployed at a plurality of key positions of the power grid, including access points of substations, power transmission and distribution line nodes, load centers and important equipment, so as to realize comprehensive monitoring of the running state of the power grid.

First, the voltage sensor and the current sensor are installed at the high-voltage bus of the transformer substation, the intermediate node of the power transmission and distribution line, and the important load access point. The voltage sensor can collect instantaneous values of node voltage in real time and is used for monitoring stability of grid voltage, and the current sensor is responsible for capturing amplitude and direction information of node current and is used for evaluating current distribution and load conditions. The sensor transmits the acquired data to the integrated data collection unit of the module via a standard communication protocol such as MODBUS or IEC 61850.

Secondly, temperature sensors are installed on critical contacts of the line, the housing of the high load device and the transformer of the power grid for monitoring the temperature change of the device. Temperature data is particularly important because temperature rise may indicate line overload or equipment anomalies. The sensor transmits data to the central processing unit of the multi-level sensing module through a wireless communication module (such as ZigBee or LoRa).

For frequency detection, a high-precision frequency acquisition device is integrated in the module and is installed on the low-voltage side of a main transformer substation, so that the operation frequency of a power grid can be monitored at a high sampling rate (for example, 100 times per second), and frequency offset caused by load fluctuation or faults is captured. In addition, the device has an anti-interference function, and can maintain high-precision measurement in a strong electromagnetic interference environment.

The multi-level sensing module 101 further includes a data integration and encapsulation unit for summarizing, formatting and packaging the data from the different types of sensors to form a standardized data packet. The unit ensures the consistency of the time stamps of all acquired data through a time synchronization mechanism (such as GPS-based clock synchronization), thereby improving the accuracy of data analysis.

Finally, the module is equipped with reliable data transfer functions. The integrated communication module supports various communication modes, including wired (e.g., ethernet or optical fiber) and wireless (e.g., LTE or 5G), to ensure that the collected data can be quickly and stably transferred to the data processing module 102. In order to improve the robustness of data transmission, the module is also provided with a data buffering and retransmission mechanism, so that the data can be temporarily stored when the communication is interrupted and retransmitted after recovery, and the data loss is avoided.

Through the design, the multi-level sensing module 101 realizes multi-dimensional monitoring and efficient data transmission of the running state of the power grid, and reliable data support is provided for subsequent data processing and fault diagnosis.

Further, the multi-level sensing module is specifically configured to:

The multi-level sensing module fully considers the operation monitoring requirements of different positions in the power grid, and realizes real-time acquisition and transmission of operation data such as voltage, current, temperature, frequency and the like by deploying multi-type sensors in a transformer substation, a power transmission and distribution line, a load end and other key areas, thereby providing comprehensive basic data support for monitoring the operation state of the power grid and fault diagnosis.

At the substation, transmission and distribution lines and load ends, voltage sensors and current sensors are reasonably deployed to cover the main nodes and paths of the power grid. The voltage sensor is specially used for detecting the voltage instantaneous value of the node, and the design accuracy can meet the high sensitivity requirement on voltage fluctuation in the operation of the power grid. These sensors are typically connected to the critical points of the high voltage bus and transmission line and are capable of capturing rapid drops in voltage or other abnormal changes when a fault occurs. The current sensor is arranged at the inlet of the power transmission and distribution line and the load equipment and has the capability of monitoring the current amplitude and direction. These sensors provide critical data support for fault localization and load analysis by detecting the flow characteristics of the current.

At high load equipment and critical line locations, temperature sensors are used to monitor the temperature rise of the equipment enclosure or line. The primary goal of temperature monitoring is to prevent overheating problems due to overload or equipment failure, which may lead to electrical fires or equipment damage. The installation point of the temperature sensor selects a heating part of equipment as a core, such as a transformer oil tank, a switch cabinet shell and a high-voltage line connection point. The temperature data is sampled at regular time intervals and transmitted to the data processing unit via the wireless communication module. The wireless communication module supports multiple protocols (such as ZigBee, loRa or LTE) and can keep stable data transmission in a complex electromagnetic environment.

At the main substation, frequency acquisition means are integrated into the system for continuously sampling the system frequency in operation of the grid. These devices have high sampling rate and high accuracy characteristics and are capable of capturing frequency shifts caused by load fluctuations or faults. For example, when a short circuit or a large-scale load switch occurs, the frequency acquisition device can instantly detect the change of the system frequency and transmit the change as an abnormal initial signal to a subsequent data processing module. In order to ensure the synchronism of the collected frequency data, the frequency collecting device is generally internally provided with a high-precision clock source and supports time alignment with the data of other sensors.

All sensor data are preliminarily packaged through an integrated processing unit of the multi-level sensing module after being collected, so that data packets with standardized structures are formed, and the data packets comprise information such as measured values of voltage, current, temperature and frequency, time stamps, equipment identifiers and the like. The integrated processing unit also has a data integrity checking function, and can perform error checking on the acquired data and request retransmission when necessary so as to ensure the accuracy of the data. After the data package is completed, the data are sent to a data processing module in a wired communication or wireless communication mode for further analysis and diagnosis.

Through the implementation steps, the multi-level sensing module can meet the diversified requirements of monitoring the running state of the power grid, has high-efficiency and reliable monitoring and data transmission functions, and provides a solid foundation for the running of the whole power grid real-time fault detection and automatic repair system.

The data processing module 102 is configured to receive the operation data provided by the multi-level sensing module, analyze the operation data through an embedded intelligent algorithm to identify a potential fault, and generate a fault diagnosis result when the fault is detected, where the fault diagnosis result includes a fault type, a fault location and an influence range.

The data processing module 102 is one of the core components of the power grid real-time fault detection and automatic repair system, and has the functions of receiving the operation data from the multi-level sensing module 101, analyzing the operation data by using an embedded intelligent algorithm, thereby identifying potential faults and generating diagnosis results of fault types, fault positions and influence ranges. The design of the module covers the whole process of data receiving, preprocessing, analyzing, diagnosing and outputting, and all operations aim to be efficient and automatic.

Upon receiving the operational data, the data processing module interfaces with the multi-level awareness module via standardized communication interfaces that support a variety of communication protocols, such as MODBUS, IEC 61850, or TCP/IP based communication protocols. The module is internally configured with a receiving unit capable of processing data streams from multiple sensors simultaneously, including parameters such as voltage, current, frequency, and temperature. In order to ensure the integrity of the received data, the receiving unit has an error correction mechanism, and can detect and repair bit errors in the data transmission process, and simultaneously support a temporary data caching function so as to cope with short-time communication interruption.

The operation data enters the data processing module and then is subjected to standardization and denoising processing through the preprocessing unit. First, the module normalizes each data stream by its physical quantity unit to eliminate analysis bias due to dimensional differences. For example, for voltage and current data, the preprocessing unit will convert it to a percentage form depending on the rated parameters of the device. And secondly, the module utilizes a sliding window filtering technology and a threshold value removing algorithm to clean high-frequency noise and abnormal values in the data, so that the accuracy of subsequent analysis is ensured. Meanwhile, the preprocessing unit distributes uniform time stamps for data from different sensors through a time synchronization mechanism so as to ensure the accuracy of time correlation in multidimensional data analysis.

After the preprocessing is completed, the module performs feature extraction on the cleaned data, which is the basis of fault diagnosis. The characteristic extraction unit combines the characteristics of the power grid operation data to design a set of multi-level characteristic analysis logic. For example, for current data, the module calculates the waveform change rate and the current difference between adjacent nodes, which can effectively reflect the unbalanced condition of the grid load, and for voltage data, the module extracts the unbalance degree and the fluctuation amplitude for identifying the voltage abnormality region. In addition, the module analyzes the dynamic change mode of the frequency data, and provides support for subsequent diagnosis by capturing the changes of the main frequency offset and the harmonic components.

Fault diagnosis is a core function of the data processing module. The module is embedded with a pre-trained intelligent algorithm which is based on a Support Vector Machine (SVM) model and dynamically optimized in combination with load data updated in real time. The algorithm receives the extracted feature data as input and generates a fault diagnosis result through classification and regression analysis. In particular, the classification section is used to identify the type of fault, such as a short circuit, overload or equipment anomaly, and the regression analysis calculates the location of the fault, such as a particular line or node, based on the characteristic data. In order to improve the accuracy of diagnosis, the algorithm dynamically adjusts the boundaries of the classification model by combining historical sample data and current load distribution, thereby adapting to the change of the real-time running environment.

After the fault diagnosis result is generated, the data processing module packages and outputs the result, including key information such as fault type, fault position, influence range and the like. The module is equipped with standardized output units supporting data encapsulation in multiple formats, such as JSON, XML or binary data streams. The output unit can send the diagnostic result to the automated repair module 103 via the high-speed communication interface to trigger the repair operation, while delivering the result to the monitoring and logging module 104 for recording and display.

The data processing module 102 can efficiently process mass data in the operation of the power grid, can accurately identify and position faults, and provides powerful technical support for the stable operation of the power grid.

The intelligent algorithm adopted in the data processing module realizes comprehensive analysis, fault diagnosis and positioning of the running state of the power grid and accurate evaluation of the fault influence range through multi-step coordinated analysis and calculation flow. The implementation of the algorithm covers a plurality of stages of data preprocessing, feature extraction, fault diagnosis, physical structure backtracking and influence range analysis, and the design of each stage ensures the logic property and technical feasibility of the algorithm.

In the first stage, the algorithm receives the original operation data from the multi-level sensing module, including signals such as voltage, current, frequency and temperature. These signals are often subject to high frequency noise, so the algorithm first decomposes the data by wavelet transformation. The wavelet transform can divide the operating data into different frequency bands, where the high frequency components contain noise information, while the low frequency components preserve the main grid operating characteristics. By removing the high frequency components and retaining the low frequency components, the algorithm generates a set of time series data reflecting the dynamics of the grid. These data are then processed in segments, each time series being analyzed using a sliding window technique to capture local changes in operating conditions.

The results of the sliding window analysis are used for feature extraction, which is the basis for fault diagnosis. The characteristic extraction stage focuses on three core indexes, namely the change rate of a current waveform, the voltage unbalance degree and the frequency offset characteristic. The current waveform change rate is obtained by calculating the current change amplitude of adjacent time points and is used for identifying current abnormal fluctuation, the voltage unbalance degree is calculated by comparing voltage differences among different phases of a power grid and reflects the stability of voltage distribution, and the frequency deviation characteristic captures the deviation and the change trend of the running frequency of the power grid and can effectively indicate load fluctuation or fault influence.

The extracted characteristic data are input into a pre-trained support vector machine classifier to perform fault diagnosis. The support vector machine uses the historical sample data to perform initial training, including various fault types and corresponding feature distribution, and simultaneously optimizes classification boundaries by combining real-time load data, so that diagnosis accuracy and real-time performance are improved. The output of the classifier is a diagnostic result of the type of fault, such as a short circuit, overload, line break, or equipment anomaly. In addition, the diagnosis results can mark abnormal operation parameters related to the fault type, such as current peak value or frequency abnormal value at a specific time point, so as to provide basis for subsequent analysis.

Based on the diagnosis result, the algorithm performs layer-by-layer retrospective analysis on the physical structure of the power grid to determine the source of the voltage or current variation. The process utilizes the topological information of the power grid and the propagation path of the abnormal current to trace back gradually from the affected load equipment to the power transmission and distribution line until an initial node of abnormal change is found. By combining the operating parameters with the historical data, the algorithm further analyzes the source node for abnormal behavior, such as sudden current rises or voltage dips, and ultimately determines the location of the faulty equipment or line. The result of this stage is the generation of an accurate fault location report that includes the identity of the faulty equipment or line.

After determining the fault location, the algorithm further evaluates the overall impact range of the fault on the grid operating state. Through step-by-step propagation analysis, an algorithm simulates the diffusion process of abnormal current or voltage, and dynamically calculates the influence range by combining the operation characteristics of equipment and circuits. The result of each propagation updates the operating state changes of the related equipment, including current load increase, voltage drop increase or frequency offset diffusion, etc. When the propagated effect weakens to a predetermined negligible range, the algorithm stops the propagation calculation and generates a detailed effect report. The report lists the identity of all affected devices, the specific values of the changes in the operating parameters, and the geographic or topological distribution of the scope of influence.

Through the collaborative work of the stages, the intelligent algorithm in the data processing module can realize the full-flow automatic processing from the preliminary analysis of the operation data to the fault positioning and the influence evaluation, and provides technical support for the rapid repair and the comprehensive evaluation of the power grid faults.

wherein t represents sampling time and is a time coordinate of power grid operation data;

j represents a scale parameter, and controls the balance of the time resolution and the frequency resolution of the wavelet basis function;

k represents a time shift parameter, determining the position of the wavelet on the time axis.

ψ(t)=exp(-αt²)cos(2πβt) (3)

φ(t)=exp(-αt²) (4)

Wherein α is a frequency offset adjustment parameter, used for controlling the time domain expansion degree of the wavelet function, and usually selecting a smaller value according to the frequency characteristic of the power grid signal to enhance the time resolution;

beta is a waveform adjustment parameter, defines the central frequency of a wavelet function, and generally selects a value consistent with the main frequency characteristic of a power grid signal;

t represents a sampling time;

D _j (t) represents a convolution operation of the operational data x (t) with the wavelet basis function ψ _j,k (t) for capturing the high frequency varying part of the signal.

A _j (t) is used to calculate the convolution of the running data x (t) with the scaling function phi _j,k (t), extracting the low frequency characteristics of the signal. The low frequency characteristic of the signal is extracted.

wherein η is a scaling factor set based on the statistical properties of the grid operation data. Typically, by statistically analyzing the historical operating data, a value is selected that is effective to distinguish between noise and signal, e.g., η is preferably in the range of 1 to 3.

N is the number of sampling points;

Further, the performing sliding window time sequence analysis on the low frequency component, extracting feature data reflecting the running state of the power grid, and generating the power grid feature data includes:

And respectively carrying out wavelet decomposition on the operation data of the current, the voltage and the frequency to generate corresponding low-frequency components A _current(t)、A_voltage (t) and A _frequency (t), carrying out time sequence analysis on each low-frequency component by adopting a sliding window technology, defining the sliding window length W and the step length S, and extracting the following various characteristics from different windows.

The low frequency components a _current(t)、A_voltage (t) and a _frequency (t) herein can be obtained by the methods of the above formulas (1) to (7). The sliding window technique effectively captures local variation characteristics by processing the signal in segments. The length W of the sliding window determines the time span of each analysis, typically set according to the characteristic period of the grid signal, e.g. W may be selected in the range of 0.1 to 1 second at a grid frequency of 50 Hz. The step size S defines the number of steps in the window sliding interval, with smaller steps increasing the resolution and larger steps reducing the computational burden, typically 10% to 20% of W.

The current waveform change rate R _current is calculated according to the following formula (8):

Wherein N is the number of data segments in the window, n=w/S, m _i is the fit slope of the ith data segment in the window, m _i-1 is the fit slope of the ith-1 data segment in the window, and m _i is calculated by adopting the following formula (9):

Wherein W _i is the ith sliding window, its time range is [ t _i,t_i +W ], wherein W is the window length, t _i is the start time of the ith sliding window, t is the sampling time point; Is the mean value of the time points in the window, and is calculated by the following formula (10):

Wherein n is the number of sampling points in the window; Is the mean value of the low frequency components in the window, and is calculated by the following formula (11):

Wherein n is the number of sampling points in the window;

the voltage unbalance U _unbalance is calculated according to the following formula (12):

wherein t _k represents the starting time of the kth sliding window, W is the length of the sliding window, S is the step length, the numerator part calculates the sum of the voltage differences between adjacent sampling points, and the denominator part normalizes the result by using the maximum value and the minimum value of the signal, thereby facilitating the comparison of signals in different amplitude ranges.

The frequency offset characteristic F _offset is calculated according to the following formula (13):

Wherein ω _normal is the grid normal frequency (typically 50Hz or 60 Hz), ω _inst (t) is the instantaneous frequency at time t, calculated using equation (14) as follows:

Wherein H represents a Hilbert-Huang transform, arg is a phase angle calculation function;

And integrating the calculated current waveform change rate R _current, the voltage unbalance U _unbalance and the frequency offset characteristic F _offet to generate power grid characteristic data.

Further, the inputting the feature data into a pre-trained support vector machine classifier, dynamically optimizing classification boundaries based on historical samples and real-time load data, and generating a diagnosis result of a fault type includes:

the method comprises the steps of receiving characteristic data of the running state of a power grid, wherein the characteristic data comprise current waveform change rate, voltage unbalance degree and frequency offset characteristics;

Training a support vector machine classifier by using a historical sample data set, wherein the historical sample data set comprises a fault type label and corresponding characteristic data;

dynamically adjusting the decision boundary of the classifier in combination with the real-time load data, and optimizing the classification model by updating a weighted distance formula among support vectors, wherein the optimization target is the following formula (15):

Wherein w is the weight vector of the classifier, ζ _i is the relaxation variable of the ith sample, n is the number of samples, β is the load sensitive weight coefficient, load _i is the load data of the ith sample, and load represents the load data of all samples.

The power grid real-time fault detection and automatic repair system utilizes a support vector machine classifier to diagnose fault types, combines historical sample data and real-time load data, and generates accurate diagnosis results through dynamic optimization classification boundaries.

The input of the support vector machine classifier is characteristic data of the running state of the power grid, wherein the characteristic data comprise the change rate of a current waveform, the unbalance degree of voltage and the frequency offset characteristic. Since these feature data are different in dimension, for example, the current waveform change rate is expressed in amplitude change, and the frequency offset characteristic is in hertz, normalization processing is required for all the feature data before inputting into the classifier. Normalization ensures that the range of all features is the same by scaling the value of each feature by its maximum or standard deviation, typically normalized to within the [0,1] or [ -1,1] range, thereby avoiding the effect of dimensional differences on classifier performance.

The training process of the support vector machine classifier is based on a pre-prepared historical sample dataset. The historical sample dataset includes a plurality of samples of known fault types, each sample containing corresponding characteristic data and tags. The labels describe the types of faults such as short circuits, overloads, line breaks, or equipment anomalies. Through the sample data, the classifier can learn the characteristic modes corresponding to various faults. During the training process, the classifier separates different types of fault data by optimizing the classification boundary so as to accurately classify unknown faults during operation.

The optimization objective of the classifier is based on the following formula:

where w represents the weight vector of the classifier, which determines the location and direction of the classification boundary. The classification boundary is a hyperplane in a high-dimensional space that can effectively separate fault types. First term of optimization objective The complexity of the classification boundary is controlled, and the classifier is prevented from being overfitted by the smaller weight value w.

The relaxation variable ζ _i in the second term is a tolerance to classification errors. For samples that are difficult to classify, the value of ζ _i will be large, allowing some error in the classifier on these samples. The weight C determines the weight of the error term, a larger value of C tends to strictly reduce the classification error but may lead to overfitting, and a smaller value of C allows a larger classification error but can improve the generalization capability of the model.

The formula also comprises a load sensitive weight adjustment termWhere β is a coefficient of sensitivity that controls the extent to which real-time loading affects the adjustment of the boundaries of the classification. The coefficients are typically selected by experimental data optimization, with values ranging typically between 0.1 and 1. The real-time load data load _i is the load information of a particular sample in operation, and max (load) is the maximum value of all sample loads for normalization processing. By introducing the term, the classifier can give higher weight to the samples with larger loads, so that the classification boundary is dynamically adjusted, and the classifier is more suitable for the load characteristics in a real-time running environment.

In the actual optimization process, the support vector machine minimizes the objective function using gradient descent or other optimization algorithms. With the progress of optimization, the classification boundary can be gradually adjusted, and finally the effect of maximally distinguishing different fault types is achieved. After the optimization is completed, the classifier outputs the diagnosis result of the fault type, including the label of the fault type and the confidence score.

Through the design, the classifier can be dynamically adjusted by combining the operation data of the power grid in real time, so that the accurate classification and efficient diagnosis of the fault types of the complex power grid are realized. Parameters C, beta and normalization range can be flexibly adjusted according to load characteristics, historical sample data distribution and real-time operation requirements of an actual power grid so as to optimize the performance of the classifier and adapt to different application scenes.

Furthermore, the analyzing the source of the voltage or current variation by tracing back the physical structure of the power grid layer by layer based on the abnormal current propagation path according to the diagnosis result of the fault type and the marked abnormal operation parameters comprises:

based on the fault type diagnosis result and the abnormal operation parameters, identifying an initial node associated with abnormal data in the power grid physical structure;

The physical structure of the power grid is traced back layer by layer, an upstream node and a downstream node or a line connected with the initial node are sequentially checked along an abnormal current propagation path, and the propagation direction of current or voltage abnormality is judged by comparing the voltage variation difference value and the current imbalance characteristic of the adjacent nodes;

and combining historical operation data, a topological structure and a real-time load state in each layer of backtracking, identifying a main path of abnormal propagation, finally determining an initial source of current or voltage change, and marking the initial source as a suspected fault point for subsequent analysis.

According to the embodiment, the physical structure of the power grid is traced back layer by layer, the source of voltage or current change is analyzed by taking the abnormal current propagation path as a basis, and the initial source for causing abnormality is determined step by combining the diagnosis result of the fault type and the marked abnormal operation parameters, so that a key basis is provided for fault positioning. The specific implementation procedure is described in detail below.

First, the system identifies an initial node in the grid physical structure that is associated with the anomaly data based on the fault diagnosis results and the marked anomaly operational parameters. An initial node refers to a grid node or device directly associated with an abnormal operating condition in fault type analysis, such as an access point for a particular load side, substation, or transmission and distribution line. When the initial node is identified, the system utilizes real-time monitoring data (such as current fluctuation and voltage abnormal value) to be combined with the power grid topological structure, and relevant nodes are locked through parameter matching and logic deduction. For each initial node, the system not only records the abnormal parameter value of the initial node, but also synchronously extracts the connection relation of the upstream node and the downstream node of the initial node, and the information is used as the basis for judging the abnormal propagation path in the subsequent backtracking analysis.

After determining the initial node, the system starts a layer-by-layer backtracking process of the physical structure of the power grid. The backtracking process is based on topology information of the power grid, and starts from an initial node, and gradually tracks to the source of the power grid along the propagation path of the abnormal current. In each layer backtracking, the system analyzes the voltage variation and current characteristics of the upstream and downstream nodes or lines directly connected to the current node. Specifically, by comparing the voltage variation differences between adjacent nodes, the system can determine the propagation direction of the voltage anomaly. Meanwhile, whether the current abnormality is continued to the node is further verified by analyzing the current unbalance characteristics (such as abrupt change of the current amplitude or phase deviation) of the adjacent lines.

In the backtracking process, the system also combines historical operation data, real-time load states and topological structure characteristics to optimize judgment of the abnormal propagation path. The historical operating data provides patterns of behavior of the node or line in normal and abnormal conditions, such as load fluctuation ranges, line impedance characteristics, and the like. The real-time load status provides a dynamic reference for the current load distribution and operating conditions, such as whether a node is overloaded or has a power imbalance. By integrating the information, the system can more accurately judge the direction and the main path of the abnormal propagation, and avoid the erroneous judgment caused by the fluctuation of a single parameter.

Finally, the system locks the initial source of current or voltage variation according to the abnormal propagation path. The source node is the primary location of the anomaly, and may be the point of anomaly due to equipment failure, short circuit, or other physical problems. When the source node is identified, the system marks it as a suspected fault point and appends information including its physical location (e.g., substation or line number), primary anomaly parameter values (e.g., current peak or voltage drop values), and associated upstream and downstream nodes. This information provides critical support for subsequent further analysis and repair operations.

Through the implementation steps, the method not only can accurately trace back the propagation path of the current or voltage abnormality, but also can accurately identify the source fault point from the complex power grid structure, thereby providing technical support for quick repair and reliable operation of the power grid.

Further, the determining the specific location of the fault according to the analysis result of the source of the voltage or current variation and the operation characteristics of the equipment and the line includes:

Analyzing the marked suspected fault points and the operation parameters of peripheral equipment thereof, and checking the load current change rate of the equipment, the voltage drop amplitude of a circuit and the time sequence characteristic of current unbalance;

The operation characteristics of the equipment and the line are combined, wherein the operation characteristics comprise rated load capacity of the equipment, impedance characteristics of the line and historical fault modes, whether suspected fault points accord with actual fault characteristics is comprehensively evaluated, and most probable fault points or fault lines are screened out;

and confirming the final fault point position through matching with the power grid topology information.

According to the source analysis result of voltage or current variation and combining the operation characteristics of equipment and lines, the specific position of fault occurrence is determined, and an accurate method for positioning the power grid faults is provided. According to the process, through comprehensive analysis of suspected fault points and peripheral equipment thereof, specific positions of faults are confirmed step by combining operation characteristics and topology information of a power grid, and accurate basis is provided for subsequent repair operation.

In the first stage of fault location, the system performs in-depth analysis on the marked suspected fault points and the operation parameters of peripheral equipment thereof. These operating parameters include the rate of change of the load current of the device, the magnitude of the voltage drop across the line, and the timing characteristics of the current imbalance. The load current rate of change may reflect the stability of the operation of the device, e.g., a sudden change in current over a short period of time may indicate that the device is being abnormally affected. The voltage drop amplitude of the line is a key indicator for measuring the health condition of the power transmission line, and the significant voltage drop may be caused by the increase of the impedance of the line or short circuit. In addition, analysis of the timing characteristics of the current imbalance may reveal whether there are asymmetric loading or phase failure issues in the three-phase system, which are typically accompanied by faults.

In the second stage, the system comprehensively evaluates the suspected fault points by combining the operation characteristics of the equipment and the lines. The operating characteristics include rated load capacity of the device, impedance characteristics of the line, and historical failure modes. For example, rated load capability is used to determine whether the current load of the device is outside of a design range, and line impedance characteristics may help evaluate whether the voltage drop amplitude is consistent with normal operation. By introducing historical failure mode data, the system can identify whether the current situation coincides with a past known failure mode, thereby further verifying the reliability of the suspected failure point. Based on the analysis results, the system prioritizes all suspected fault points and screens out the most probable fault points or fault lines.

Finally, by matching with the topology information of the power grid, the system confirms the final fault point position. The topology information includes the connection relationship between the devices and lines in the power grid, and the geographic or physical location of each node. The system uses topology information to verify whether the screened fault points conform to the actual current or voltage propagation paths and checks whether the running states of the upstream and downstream devices or circuits are consistent with the fault characteristics. By the method, error marks caused by abnormal data or accidental fluctuation can be effectively eliminated, and the accuracy of a final positioning result is ensured.

The whole fault positioning process technically realizes multi-level optimization from data analysis to comprehensive judgment and then to final confirmation. The design of the system ensures that the location of the fault point is not only dependent on a single parameter, but provides a highly accurate fault diagnosis result through multi-dimensional data verification and logic derivation.

And the automatic repair module 103 is used for executing repair operation according to the fault diagnosis result generated by the data processing module, wherein the repair operation comprises fault region isolation, grid topological structure reconstruction and normal power supply restoration.

The core of the automated repair module 103 is to efficiently and automatically handle grid faults, the operation of which includes three main phases, fault region isolation, topology reconstruction and power restoration.

First, during the fault zone isolation phase, the module initiates an isolation control operation using the fault diagnostic results received from the data processing module 102. A dynamically updated power grid topology database is maintained in the system, and all lines, nodes and connection relations of the lines, the nodes are recorded. Upon receiving a specific fault location, the isolation control unit extracts a list of switching devices directly connected to the faulty node or line from the database and establishes a control channel with these devices based on a real-time communication protocol (e.g., IEC 61850). The isolation instruction is sent to the target device in a standard format through the channel, and the instruction content includes a unique identification of the switch, an execution action (such as disconnection), a trigger condition, and a confirmation mechanism. To ensure the reliability of the isolation operation, the module monitors the status feedback signal of the switching device in real time. If the confirmation signal of successful execution is not received within the set time, the module automatically sends a repeat instruction or alarm information to the standby control path.

In the topology reconstruction stage, the module relies on a built-in topology optimization engine that calculates an optimal power supply path based on a depth search algorithm and load balancing rules. Firstly, the module extracts all node and line information of a non-fault area from a power grid topology database, and calculates the residual transmission capacity of each line by combining real-time load data. Second, the engine views the grid as a directed graph, where the nodes represent substations or load devices and the edges represent transmission lines. The method is characterized in that the edge weight is given by a dynamic weighting method, the residual capacity of a line is used as a main parameter, and a shortest path or an optimal path from a power supply to a load is found by utilizing depth-first search or Dijkstra algorithm. The finally generated path scheme is sent to the control unit in the form of instructions, and the instruction content comprises switching equipment, load scheduling information and safety verification parameters of the lines, wherein each line needs to be connected.

The power supply recovery stage is completed by a power supply recovery unit, and the operation flow is refined into a gradual power supply recovery mechanism. First, the module determines the priority of power restoration based on the topology optimization results, typically to preferentially restore lines to which critical load devices (e.g., hospitals or communication base stations) are connected. The specific implementation mode of the recovery operation is that the line is activated section by section, and the whole running state of the power grid is monitored after each activation, including whether the voltage fluctuation range, the frequency stability and the current distribution exceed the set threshold value. If an anomaly is detected during any one of the recovery procedures, the module immediately pauses the subsequent operation and notifies the monitoring module to re-evaluate the repair scheme. To prevent grid instability due to sudden load increases, the power restoration process introduces a load segment activation strategy, e.g., to step up the load over a preset time interval.

The automatic repair module 103 not only can quickly isolate a fault area and dynamically reconstruct the power grid topology, but also can gradually recover power supply in a stable and safe mode, thereby remarkably improving the reliability and the operation efficiency of the power grid.

Still further, the automated repair module is specifically configured to:

The automatic repair module realizes isolation of power grid faults, dynamic adjustment of topological structures and power supply recovery of non-fault areas through a series of coordinated operations, and ensures that the power grid can run rapidly and stably. The module starts corresponding repair operation based on fault diagnosis results, reduces manual intervention through a highly-automatic process, and improves repair efficiency and accuracy.

When the data processing module generates a fault diagnosis result, the automatic repair module can immediately start the circuit breaker control unit to isolate a fault area. The module firstly analyzes the specific fault position and the influence range in the fault diagnosis result, matches the specific fault position and the influence range with a topological structure database of the power grid, and determines the line or equipment to be isolated and the serial numbers of the associated circuit breakers. The isolation operation is accomplished by the control unit sending instructions to the corresponding circuit breaker or switching device, the instructions including device identification, execution actions (e.g., disconnection), and an operation confirmation mechanism. In order to ensure the reliability of operation, the module is designed with a real-time monitoring function, and the execution condition of the instruction is verified by receiving a state feedback signal of the circuit breaker. If the target device fails to be disconnected within a specified time, the module may trigger a standby scheme, such as resending the isolation instruction or reporting an exception to the monitoring module.

After the isolation of the fault area is completed, the module enters a power grid topology adjustment stage. At this point, the topology optimization engine, which receives the operation, calculates a new power supply path using the non-fault region operation data and topology information acquired in real time. The optimization engine selects the optimal backup power source, backup line, or redistributes the existing load based on load balancing rules and path priority rules. In particular, the backup power source may include a temporarily enabled genset or unused power supply lines, while load redistribution is accomplished by dynamically adjusting the transformer output or switching the load connection points. The generated new path scheme is packaged into an instruction set and sent to the related switch equipment through a communication interface built in the module. These instructions include a list of devices that need to be turned on or switched, load adjustment parameters, and security check rules to ensure that the topology adjustment is performed safely and reliably.

After topology adjustment is completed, the automatic repair module starts a power restoration process, and the goal of this stage is to gradually restore normal power to the non-fault area. The power restoration is responsible for the power restoration unit, the operation of which is performed in a piecewise activated manner, avoiding secondary problems caused by load surges. Before each activation, the module will evaluate whether the current load capacity is sufficient to support the new power demand by monitoring the real-time operating conditions of the transformer and line. If an anomaly is detected, such as an overload or voltage fluctuation exceeding a threshold, the module may suspend the resume operation and re-evaluate the resume policy. During the power restoration process, the module can also dynamically adjust the voltage level and the output power of the transformer to adapt to the gradually increased load. Finally, after the power supply recovery of all the load areas is completed, the module can inform the monitoring module to record the detailed data of the whole repair process.

The automated repair module is also equipped with a redundant security mechanism that can quickly switch to a standby scheme in the event of isolation or topology adjustment failure. In order to achieve the above, a real-time abnormality detection and response function is designed in the module, when any link is detected to have an abnormality, the system immediately pauses the current operation and sends the abnormality to the monitoring module, and at the same time, attempts to re-plan the repair strategy or requests manual intervention.

Through the design, the automatic repair module can realize full-flow automatic operation from fault isolation to power supply recovery, and the flexibility and reliability of the automatic repair module obviously improve the risk resistance of the power grid.

The monitoring and logging module 104 is used for recording the running state of the power grid, fault diagnosis results and repair operation processes, generating a historical data report, and providing fault analysis and running performance evaluation through a user interface.

The monitoring and logging module 104 is a key part in the real-time fault detection and automatic repair system of the power grid, and has the main functions of recording the running state of the power grid, fault diagnosis results and repair operation processes, generating detailed historical data reports, and providing fault analysis and running performance assessment through a user interface. The module design combines the dual requirements of real-time performance and data analysis, and provides comprehensive support for operation management of a power grid through an efficient recording mechanism, a flexible display mode and a powerful analysis function.

The monitoring and logging module first receives real-time data and operation results through communication interfaces with the data processing module 102 and the automated repair module 103. The received data includes operating state parameters such as voltage, current, frequency, temperature, etc., type, location, area of impact of fault diagnosis, and steps and time nodes of repair operations. The module distributes accurate time information for each record through a time stamping mechanism, and ensures that data from different sources keep consistent in the time dimension.

The recording function is the basis of the module, and key events in the running of the power grid are recorded in a layering mode through a log recording unit. The first layer is a real-time recording layer, and the layer carries out original recording on all operation data and fault information, so that the integrity and traceability of the data are ensured. The second layer is an event recording layer for capturing and collating important events in operation, such as fault occurrence, time points when repair operations begin and end, and warning or anomaly information related to the operating state of the power grid. The third layer is a report generation layer that generates a formatted historical data report from the recorded data, including an operational curve, a fault summary, and a repair operation log. These reports may be exported in a common format (e.g., PDF or CSV) for long-term storage and further analysis.

In order to facilitate the real-time monitoring and interaction of users, the module is provided with a user interface unit, and the running state and the history of the power grid are presented in a visual mode. The user interface adopts a multidimensional display method and comprises a real-time monitoring view, an event list and a performance evaluation instrument panel. In the real-time monitoring view, parameters such as voltage, current, frequency and the like are displayed in a dynamic chart form, and a user can intuitively check the current running state of the power grid. All critical events are listed in the event list, and the user can click on the event to obtain detailed information, such as a topology map of the location of the fault or an execution step of the repair operation. The performance assessment dashboard provides a quantitative assessment of grid operational quality through metrics such as average repair time, frequency of failure occurrence, and system stability scores.

The monitoring and log module also has an intelligent analysis function and can carry out deep mining based on historical data. For example, the module may identify common failure modes by cluster analysis of multiple failure data and discover potential causes related to the failure by correlation analysis. In addition, the module supports simulation of past faults and repair processes through playback functions, helping users verify the response speed of the system and the effectiveness of the repair strategy.

The design of the module fully considers the data security and reliability. All log records are stored in a redundant mode and stored in a local server and a cloud server so as to prevent data loss. The module also employs a rights management mechanism to ensure that only authorized users can access sensitive data or perform data export operations.

The monitoring and logging module 104 realizes the full-flow function from data recording to advanced analysis, and provides solid support for transparent management and continuous optimization of the running state of the power grid.

Further, the monitoring and logging module is specifically configured to:

The monitoring and log module is an important component of the power grid real-time fault detection and automatic repair system and is responsible for realizing comprehensive monitoring of the running state of the power grid, accurate recording of fault data and complete tracking of the repair process, and reliable data support is provided for subsequent running evaluation and optimization. The module ensures the transparency of the running state and the systematic management of the information through the cooperative work of the state acquisition unit, the real-time recording function and the log report generation unit.

The module acquires operation data of the power grid in real time through the state acquisition unit, wherein the operation data comprise key parameters such as voltage, current, frequency and temperature. The data acquisition is based on a communication interface with the multi-level perception module, and the acquisition unit can receive the data stream at high frequency and attach an accurate time stamp to each piece of data so as to ensure the time sequence consistency of the data. In a normal running state, the module continuously stores the acquired data, and simultaneously monitors parameter changes in real time. If an anomaly, such as a significant shift in voltage or frequency, is detected, the module immediately triggers the recording function of the anomaly data. The abnormal record not only comprises abnormal parameter values, but also captures key operation states before and after the fault occurs, and generates a snapshot of the operation states. The snapshots provide a preliminary inference basis for the failure cause for the system by comprehensively analyzing the current and historical data.

In the fault repairing process, the monitoring and logging module plays a core recording function. The module tracks each operation of the automatic repair module in real time, and comprises the specific steps of isolating operation of a circuit breaker, topology adjustment and power supply recovery. At each execution of the operation, the module records the relevant point in time, the identification information of the executing device and the operation result, such as whether the isolating circuit breaker is successfully opened or whether there is a load abnormality in the power restoration. In addition, the module captures dynamic changes in grid operating parameters during the repair process to evaluate the effectiveness of the repair measures in subsequent analysis. For example, during a segment activation process for power restoration, the module may record current and voltage fluctuations during restoration of each segment of line, thereby determining whether an optimized power supply strategy is required.

After the fault is repaired, the monitoring and logging module enters a log report generating stage. The module can sort and collect all data in the whole repairing process to generate a complete log report. The report contains a detailed record of the grid operating conditions, the initial parameters and their trends of change for the occurrence of the fault, the complete flow of the repair operation and the final recovery results. In addition, the report includes detailed analysis of the cause of the fault, such as identifying the likely source of the fault by correlation analysis, or predicting the potential risk of similar faults by trend analysis. The log report is stored in a standardized format, supports various file forms such as PDF, CSV and the like, and is convenient for long-term archiving or export for other systems.

To facilitate quick decisions and long-term management for the operator, the module is equipped with a visual interface. The interface adopts an intuitive graphical display mode, and presents operation data and a repair process in the forms of a graph, a topological graph and a time axis. An operator can view the power grid state in real time through an interface, browse key fault information or replay the repair process. In addition, the interface also supports customized performance assessment functions, such as calculating average repair time or frequency of failure occurrence by analyzing log reports, helping grid managers optimize system operation strategies.

The design of the monitoring and logging module also focuses on the security and reliability of the data. The module adopts a dual redundancy storage mechanism to synchronously store data to a local server and a cloud server so as to prevent data loss caused by hardware failure or network interruption. The module is also provided with access rights control to ensure that only authorized users can view or export sensitive data.

Through the design, the monitoring and log module realizes the whole-flow management from data acquisition to fault analysis and from real-time monitoring to history archiving, and provides a solid technical guarantee for the efficient operation of the power grid.

While the application has been described in terms of preferred embodiments, it is not intended to be limiting, but rather, it will be apparent to those skilled in the art that various changes and modifications can be made herein without departing from the spirit and scope of the application as defined by the appended claims.

Claims

1. A real-time fault detection and automatic repair system for a power grid, comprising:

2. The system for real-time fault detection and automatic repair of a power grid according to claim 1, wherein the multi-level sensing module is specifically configured to:

3. The system for real-time fault detection and automatic repair of a power grid according to claim 1, wherein the automatic repair module is specifically configured to:

4. The system for real-time fault detection and automatic repair of a power grid according to claim 1, wherein said monitoring and logging module is specifically configured to:

5. The system for detecting and automatically repairing power grid faults in real time according to claim 1, wherein the intelligent algorithm adopted by the data processing module is realized by the following steps:

6. The system for real-time fault detection and automatic repair of a power grid according to claim 5, wherein said decomposing the operation data collected by the multi-level sensing module by wavelet transform and extracting the low frequency component by removing the high frequency noise comprises:

ψ(t)=exp(-αt²)cos(2πβt) (3)

φ(t)=exp(-αt²) (4)

7. The system for real-time fault detection and automatic repair of a power grid according to claim 6, wherein said performing sliding window time series analysis on the low frequency component extracts feature data reflecting an operation state of the power grid, and generates power grid feature data, comprising:

Performing wavelet decomposition on the running data of current, voltage and frequency respectively to generate corresponding low-frequency components A _current(t)、A_voltage (t) and A _frequency (t), performing time sequence analysis on each low-frequency component by adopting a sliding window technology, defining the length W and the step length S of a sliding window, and extracting the following characteristics from different windows:

Wherein n is the number of sampling points in the window;

Wherein t _k represents the starting time of the kth sliding window, W is the length of the sliding window, S is the step length;

Wherein ω _normal is the grid normal frequency, ω _inst (t) is the instantaneous frequency of time t, calculated using equation (14) as follows:

8. The system for real-time fault detection and automatic repair of a power grid of claim 5, wherein said inputting the feature data into a pre-trained support vector machine classifier dynamically optimizes classification boundaries based on historical samples and real-time load data, generating a diagnostic result of a fault type, comprising:

9. The system for real-time fault detection and automatic repair of a power grid according to claim 5, wherein the analyzing the source of the voltage or current variation based on the abnormal current propagation path by tracing back the physical structure of the power grid layer by layer according to the diagnosis result of the fault type and the marked abnormal operation parameters comprises:

and combining historical operation data, a topological structure and a real-time load state in each layer of backtracking, identifying a main path of abnormal propagation, finally determining an initial source of current or voltage change, and marking the initial source as a suspected fault point.

10. The system for real-time fault detection and automatic repair of electrical network according to claim 9, wherein said determining the specific location of fault occurrence based on the analysis of the source of voltage or current variation and in combination with the operating characteristics of the equipment and the line comprises: