Power grid fault detection and maintenance method and system based on relay protection

Abstract

The invention discloses a power grid fault detection and maintenance method and system based on relay protection, and relates to the technical field of power grid faults, and efficient fault handling is realized in four steps: 1, innovatively constructing an online monitoring and unmanned aerial vehicle inspection data acquisition system, synchronously deploying a CAIO and EPA collaborative autonomous network, and guaranteeing real-time data interaction; in the second step, a multi-dimensional AI evaluation model is innovatively adopted, fault risk grades are divided in combination with electrical quantities, equipment states and environmental data, and CAIO digital twinning super-real-time simulation is matched, so that a fault influence range is accurately pre-judged, and a disposal decision is dynamically output; 3, innovatively positioning a fault node through multi-EPA data cross validation and a traveling wave algorithm, screening an optimal EPA cut-off point, and customizing a differentiated reverse compensation strategy; and 4, after the arc is extinguished, an equipment exclusive maintenance scheme is innovatively generated, so that the intelligentization and precision of the whole fault handling process are realized, and the power grid fault handling efficiency and reliability are greatly improved.

Description

Technical Field

[0001] This invention relates to the field of power grid fault technology, and specifically to a method and system for power grid fault detection and maintenance based on relay protection. Background Technology

[0002] In the current power grid operation environment, the power system is continuously expanding in scale and becoming increasingly complex in structure. Various equipment are operating under high load and multiple conditions for a long time, and the probability of faults and potential risks are constantly rising. The level of intelligence and automation in emergency response is low, often requiring a large amount of manual intervention, which is not only time-consuming and labor-intensive, but may also cause more serious power grid problems due to human error. Therefore, there is a need for a power grid fault detection and maintenance method and system based on relay protection.

[0003] Existing technology, such as the invention patent application with publication number CN110912271A, discloses an intelligent monitoring and transmission system for power grid faults, belonging to the field of power grid monitoring technology. It includes a central processing unit (CPU), a signal acquisition module, and an alarm module. The signal acquisition module mainly consists of a relay protection module and a waveform recording sensor. The output terminal of the waveform recording sensor is electrically connected to the input terminal of the CPU. The waveform recording sensor detects real-time values ​​of electrical quantities such as voltage and current in the power grid line and transmits them to the CPU. The CPU retrieves standard values ​​of electrical quantities such as voltage and current stored in the data storage module and sends them along with the real-time values ​​to the data comparison module for comparison. A significant deviation between the real-time value and the standard value indicates a fault. The alarm module then triggers an audible and visual alarm. The relay protection module disconnects the power grid line to prevent leakage. The GPS positioning module locates the faulty power grid line, enabling timely monitoring and repair of faulty power grid lines.

[0004] Regarding the above-mentioned solutions, the applicant of this invention has discovered that the above-mentioned technology has at least the following technical problems: 1. Existing technologies have a single dimension for sensing equipment status, mostly limited to electrical quantity data, lacking effective monitoring of the equipment's physical structure and external environment. Furthermore, non-electrical data such as drone inspections and online monitoring data are often isolated and cannot be integrated for analysis. At the same time, communication protocols between devices are fragmented, lacking a unified collaborative communication network. Data interaction latency between edge sensing units and cloud systems, as well as between edge units, is high, making it difficult to form an efficient response mechanism of "local autonomous decision-making + global collaborative control," and easily missing the window for initial fault handling.

[0005] 2. Existing technologies have significant flaws in their judgment logic. Risk assessments often rely on single electrical quantity thresholds, failing to consider multi-dimensional indicators such as equipment service life, historical fault records, and environmental impact. This makes it impossible to differentiate risk levels based on the characteristics of different equipment, leading to frequent misjudgments and omissions. Furthermore, the parameters of digital twin simulation models are mostly statically preset and are not dynamically updated with real-time changes in the power grid topology and fluctuations in renewable energy output. The simulation results of the fault impact range deviate significantly from the actual scenario, making it difficult to accurately predict the fault propagation trend and failing to provide reliable support for the formulation of preventive measures.

[0006] 3. Existing technologies lack accuracy and fault tolerance. Fault location often relies on single-source data, failing to improve accuracy through multi-unit cross-validation, and does not consider the impact of line parameter errors and communication delays on the location results. This often results in excessive location deviations and an inability to accurately pinpoint the fault node. Furthermore, the reverse compensation current strategy parameters are fixed and lack the ability to dynamically adjust based on the real-time status of the fault node during the compensation process. When a hardware failure occurs in the edge unit performing compensation, there is no rapid switching mechanism for backup units, which can easily lead to compensation failure and arc reignition.

[0007] 4. The automation and intelligence levels of existing technologies are severely insufficient. When the reverse compensation strategy fails to meet the standards, it relies heavily on manual issuance of current cut-off and load transfer commands. The transmission and execution of commands are time-consuming, which can easily lead to further expansion of the fault. Maintenance plans are mostly generated based on general templates and are not customized according to the type and location of the faulty equipment. They also lack intuitive visual operation guidance, which can easily lead to mismatched spare parts and omissions in operation steps. Maintenance resource scheduling does not combine real-time location to optimize the route, resulting in excessive time for personnel to reach the fault point and prolonging the power outage. At the same time, modules such as data acquisition, risk assessment, and maintenance management are isolated from each other and do not achieve data interoperability with other core systems of the power grid. They also lack an algorithm iteration mechanism based on historical data, which makes it impossible to continuously improve fault handling capabilities. Summary of the Invention

[0008] To address the aforementioned technical shortcomings, the purpose of this invention is to provide a method and system for power grid fault detection and maintenance based on relay protection.

[0009] To solve the above-mentioned technical problems, the present invention adopts the following technical solution: In the first aspect, the present invention provides a power grid fault detection and maintenance method based on relay protection, including: Step 1, power grid data acquisition and networking: real-time data corresponding to each device in the target power grid is obtained through online monitoring devices and UAV inspection images. At the same time, a collaborative autonomous network of CAIO and EPA is deployed in each device in the target power grid, and a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO is built.

[0010] Step 2, Fault Risk Assessment and Determination: Based on the real-time data of each device in the target power grid, assess the fault risk level of each device in the target power grid at the current moment, and simulate the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, and then determine whether each device in the target power grid needs to start preventive measures, and output the results of whether proactive maintenance or dynamic monitoring is required.

[0011] Step 3: Fault Location and Strategy Formulation: If a device in the target power grid needs to initiate preventive measures and requires proactive maintenance, the fault node of the device in the target power grid is located and analyzed to select the optimal EPA current disconnection point for the fault node of the device in the target power grid. At the same time, the reverse compensation current strategy corresponding to the optimal EPA of the fault node of the device in the target power grid is analyzed.

[0012] Step 4: Precise Post-Fault Repair: After the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy, the repair plan corresponding to the fault node of the equipment in the target power grid is analyzed.

[0013] In a second aspect, the present invention provides a power grid fault detection and maintenance system based on relay protection, comprising: a power grid data acquisition and networking module: used to acquire real-time data corresponding to each device in the target power grid through online monitoring devices and UAV inspection images; and simultaneously, deploying a collaborative autonomous network of CAIO and EPA for each device in the target power grid, and building a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO.

[0014] Fault Risk Assessment and Judgment Module: Based on the real-time data of each device in the target power grid, it assesses the fault risk level of each device in the target power grid at the current moment, and simulates the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, thereby determining whether each device in the target power grid needs to start preventive measures, and outputs the results of whether proactive maintenance or dynamic monitoring is required.

[0015] Fault location and strategy formulation module: When a device in the target power grid needs to take preventive measures and when it needs to perform proactive maintenance, the module performs fault node location analysis on the device in the target power grid, thereby selecting the optimal EPA current disconnection point for the device in the target power grid. At the same time, it analyzes the reverse compensation current strategy corresponding to the optimal EPA for the device in the target power grid.

[0016] Precise Post-Fault Repair Module: This module is used to analyze the repair plan corresponding to the fault node of the equipment in the target power grid after the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy.

[0017] The beneficial effects of this invention are as follows: 1. In the embodiments of this invention, by combining online monitoring devices with UAV inspection images, a three-dimensional data acquisition system of "electrical quantity + equipment physical status + external environment" is constructed, which effectively eliminates the monitoring blind spots caused by relying solely on electrical quantity data under traditional technologies, and provides comprehensive data support for subsequent fault handling; at the same time, the CAIO and EPA collaborative autonomous network adopts a hybrid P2P and C / S communication protocol, and the EPA data interaction latency within the same group is ≤10ms, which not only ensures the local autonomous decision-making of edge units, but also realizes global data collaboration, solves the communication lag problem caused by traditional fragmented protocols, and creates conditions for rapid response in the early stage of faults.

[0018] 2. In this embodiment of the invention, accuracy and foresight are improved through multi-dimensional grading and ultra-real-time simulation. Based on quantitative scores of three dimensions—electrical quantity, equipment status, and environment—the risk level of equipment failure can be differentiated, avoiding the misjudgment and omission problems of traditional single-threshold judgment. CAIO calls the digital twin model and dynamically sets the simulation duration according to the fault type to accurately simulate the fault propagation path and affected area, thereby outputting intervention decisions in advance and reducing the probability of the fault spreading from a local area to the core load area.

[0019] 3. In this embodiment of the invention, the AI ​​tracing model retrieves real-time data from 3-5 EPAs and combines it with a distributed boundary judgment algorithm for direction filtering, current amplitude verification, and traveling wave distance calculation. The fault location accuracy can be reduced to the specific tower number / equipment number, significantly shortening the on-site troubleshooting time under traditional single EPA location. A reverse compensation current strategy is customized for different fault types, and the compensation effect is verified through three indicators. When the compensation fails to meet the standard, the CAIO immediately triggers the best EPA current cutoff, providing double protection to prevent fault recurrence or spread.

[0020] 4. In this embodiment of the invention, a customized spare parts list is generated based on the type and location of the faulty equipment. Combined with the AR module, three-dimensional operation guidance is pushed to reduce the problems of spare parts mismatch and operation omission caused by traditional general templates. By matching the nearest qualified team to the maintenance personnel in real time, the maintenance tools are simultaneously dispatched to the material station that can be reached within 30 minutes, reducing the time for arrival and tool preparation. During the maintenance plan analysis stage, a power restoration preparation strategy is formulated simultaneously to avoid additional power outages due to insufficient power restoration preparation after maintenance, and to accelerate the power restoration process.

[0021] 5. In this embodiment of the invention, closed-loop management of the entire process is realized. Data acquisition, risk assessment, location and handling, and maintenance and power restoration are all connected through CAIO and EPA, solving the problem of process breakage caused by the isolation of traditional technical modules. CAIO's global data storage and ultra-real-time simulation engine can accumulate fault handling data, providing support for subsequent algorithm optimization. Moreover, the collaborative autonomous network architecture can be compatible with the data acquisition and fault handling needs of new energy grid connection and new loads, laying the foundation for the intelligent upgrading of the power grid and comprehensively enhancing the stability of power grid operation and its ability to adapt to future scenarios. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart illustrating the implementation steps of the method of the present invention.

[0024] Figure 2 This is a schematic diagram of the system module connections of the present invention. Detailed Implementation

[0025] 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 embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Examples of embodiments of the present invention Figure 1 As shown, a power grid fault detection and maintenance method based on relay protection includes: Step 1, power grid data acquisition and networking: real-time data corresponding to each device in the target power grid is acquired through online monitoring devices and UAV inspection images. At the same time, a collaborative autonomous network of CAIO and EPA is deployed on each device in the target power grid to build a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO.

[0027] In a specific embodiment, the deployment of a collaborative autonomous network of CAIO and EPA in each device of the target power grid is carried out as follows: CAIO is deployed on the cloud server of each device in the target power grid, configured with a global data storage module and a super real-time simulation engine, and accesses the power grid dispatching system to obtain topology and load basic data; EPA is embedded in each smart circuit breaker of each device in the target power grid. During deployment, according to the "regional division" principle, each device in the target power grid is divided into several EPA groups, each group contains 3-5 adjacent EPAs, each group establishes a dedicated C / S communication link with CAIO, and EPAs within the same group build a P2P communication subnet.

[0028] It should be noted that CAIO is a "global decision-making and simulation hub" deployed on the target power grid cloud server. Its core functions revolve around data management, intelligent computing, and global collaboration. CAIO is equipped with a global data storage module and an ultra-real-time simulation engine. It accesses the power grid dispatch system to obtain the entire network topology, such as line connection relationships, substation locations, and basic load data, such as user electricity demand and industrial load fluctuation patterns. At the same time, it receives real-time monitoring data uploaded by each EPA, such as fault current and voltage drop amplitude, forming a database covering all dimensions of power grid operation. CAIO plays three core roles: First, it is an "intelligent assessor" that determines the risk level of equipment faults by quantifying scores based on multi-dimensional data such as electrical quantities, equipment status, and environment, avoiding the limitations of traditional single threshold judgments. Second, it is a "simulation predictor" that calls the digital twin ultra-real-time simulation engine to simulate the propagation path and impact range of different fault types, providing a basis for the formulation of preventive measures. Third, it is a "global coordinator" that issues reverse compensation current commands and fault disconnection commands to EPAs during the fault handling phase, coordinates maintenance resource scheduling during the maintenance phase, and manages power restoration preparation strategies in sync, ensuring the systematic and coherent nature of the entire power grid fault handling process.

[0029] EPA (Emerging Area Device) is a "localized sensing and execution terminal" embedded in the smart circuit breakers of various devices in the target power grid. Organized into collaborative groups according to the principle of "regional division," it serves as the "front-end touchpoint" for power grid data acquisition and fault handling. The EPA is physically embedded within the smart circuit breaker, with each EPA group containing 3-5 adjacent EPA devices, such as EPAs mounted on smart circuit breakers on adjacent towers on a certain line. EPAs within the same group achieve rapid data exchange through a P2P communication subnet. For example, in the event of a fault, they synchronously share current direction and amplitude data, with latency controlled within 10ms, ensuring high efficiency in local data sharing. Simultaneously, each EPA group needs to establish a dedicated C / S communication link with CAIO (Cybernetic Intervention Object), uploading real-time monitoring data to CAIO and receiving data from CAIO. The control commands issued (such as reverse compensation current parameters and current cut-off commands) form a closed loop of "edge perception - cloud decision-making - edge execution". The EPA plays two core roles: First, it is a "data collector", which collects electrical quantity data of the equipment in real time, such as fault current amplitude, voltage drop amplitude, and equipment status data. It also supplements the physical status information of the equipment with images from drone inspections, providing basic data for CAIO's risk assessment and fault location. Second, it is a "fault handling executor", which executes reverse compensation current output under the command of CAIO, outputting the current amplitude and phase matched according to different fault types. At the same time, it performs current cut-off operations when compensation is insufficient, quickly isolating fault nodes and preventing fault propagation. It is a key execution unit for on-site handling of power grid faults.

[0030] Through the collaborative model of "CAIO global decision-making + EPA edge execution" and combined with the P2P and C / S hybrid communication protocol, the two not only ensure the comprehensiveness and real-time nature of power grid data acquisition, but also improve the accuracy and efficiency of fault handling, forming the core architecture of the power grid collaborative autonomous network in this method.

[0031] Step 2, Fault Risk Assessment and Determination: Based on the real-time data of each device in the target power grid, assess the fault risk level of each device in the target power grid at the current moment, and simulate the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, and then determine whether each device in the target power grid needs to start preventive measures, and output the results of whether proactive maintenance or dynamic monitoring is required.

[0032] In a specific embodiment, the process of evaluating the equipment failure risk level corresponding to each device in the target power grid at the current moment is as follows: A1. According to the real-time data corresponding to each device in the target power grid, three evaluation dimensions corresponding to each device in the target power grid are obtained. The three evaluation dimensions include the electrical quantity dimension, the equipment status dimension, and the environment dimension. Then, it is evaluated whether the electrical quantity dimension, the equipment status dimension, and the environment dimension corresponding to each device in the target power grid meet the qualified conditions. If any one of the three evaluation dimensions corresponding to a certain device in the target power grid does not meet the qualified conditions, the equipment failure risk level corresponding to the device in the target power grid is evaluated.

[0033] It should be noted that through the real-time data collected by the on-line monitoring device and the UAV inspection system, and combined with the preset quantization standards, it is judged one by one: in terms of the electrical quantity dimension, the current and voltage phasor data of the device are collected through PMU and SCADA sensors. If the current mutation rate < 3% and the voltage deviation is controlled within the range of ±5%, it is judged that this dimension is qualified. Otherwise, if the current mutation rate exceeds 3% or the voltage deviation breaks through ±5%, this dimension is unqualified; in terms of the equipment status dimension, the equipment operation data collected by the on-line monitoring device at an interval of ≤ 1 minute is used. For the transformer, it is necessary to judge whether the oil temperature < 85°C, and for the circuit breaker, it is necessary to check whether the closing and opening time deviation < 20ms. If the monitoring data of the corresponding equipment all meet the above thresholds, this dimension is qualified, and if any index does not meet the standard, this dimension is unqualified; in terms of the environment dimension, based on the line and insulator images with a resolution of ≥ 4K collected by the UAV inspection system every 2 hours, environmental features such as the ice thickness of the line are extracted through an image recognition algorithm with an accuracy rate ≥ 92%. If the ice thickness < 10mm, it is judged that this dimension is qualified. If the ice thickness ≥ 10mm or environmental problems such as serious pollution of the insulator that affect the operation of the equipment are identified, this dimension is unqualified.

[0034] A2. When evaluating the equipment failure risk level corresponding to the device in the target power grid, the evaluation dimensions corresponding to the device in the target power grid that do not meet the qualified conditions are compared with the set scores for the evaluation dimensions that do not meet the qualified conditions, and the scores for the evaluation dimensions that do not meet the qualified conditions corresponding to the device in the target power grid are obtained. Then, the scores for the evaluation dimensions that do not meet the qualified conditions corresponding to the device in the target power grid are accumulated, and the accumulated value is recorded as the total score for the evaluation dimensions that do not meet the qualified conditions of the device in the target power grid. If the total score for the evaluation dimensions that do not meet the qualified conditions of the device in the target power grid is greater than 20 points, it is recorded as a potential failure. If the total score for the evaluation dimensions that do not meet the qualified conditions of the device in the target power grid is less than or equal to 20 points, it is recorded as a low-frequency failure.

[0035] It should be noted that if the electrical quantity dimension is determined to be non-compliant, the fixed base score of 40 points for that dimension will be directly matched; if the equipment status dimension is determined to be non-compliant, the fixed score of 35 points will be matched; and if the environmental dimension is determined to be non-compliant, the fixed score of 25 points will be matched. After completing the "compliance determination - score matching" for each dimension, the specific score corresponding to each non-compliant dimension of the equipment can be obtained.

[0036] In a specific embodiment, the simulation process for simulating the fault impact range of each device in the target power grid is as follows: B1. If a device in the target power grid triggers a potential fault or an emergency fault, the CAIO digital twin model is constructed based on the real-time topology and parameters of the target power grid. The operating data of the device in the target power grid is input, and the simulation duration is set. The simulation duration is 72 hours for potential faults and 1 hour for emergency faults. The fault propagation path corresponding to the device in the target power grid is output.

[0037] It should be noted that when building the model based on the real-time topology and parameters of the target power grid, the topology data is obtained from the latest line connection relationships obtained from the power grid dispatching system, including the operational status of temporary tie lines and the main wiring method of substations. The parameters are obtained from line impedance, transformer turns ratio, and real-time load power factor collected by online monitoring devices. Industrial loads are differentiated into 0.85-0.95 and residential loads into 0.9-0.98 to ensure that the model matches the actual power grid with a degree of ≥95%. When inputting equipment operation data, core data strongly correlated with the fault type are selected. For example, when the potential fault is abnormal transformer oil temperature, the oil temperature change curve and load fluctuation data for the past hour are input; when the emergency fault is a line short circuit, the current and voltage waveforms for the three seconds before the fault are input and the data is labeled. Confidence level: ≥98% when the data acquisition equipment is operating normally, and ≥90% when there is a minor fault. During the simulation, parameter verification is triggered every 10 minutes for potential faults or 10 seconds for emergency faults. If the deviation between real-time data and simulation output is >8%, or if the difference between the simulated predicted current and the actual acquired current exceeds 8%, the backup parameter set will be automatically called and generated based on historical similar operating conditions. When outputting the fault propagation path, the fault triggering sequence of each device on the path must be clearly marked, such as "short circuit of line A → tripping of circuit breaker B → voltage drop of busbar of substation C → overload of line D", and the fault threshold breakthrough time of key nodes, such as the load rate of line D reaching 105% 2 minutes and 15 seconds after the fault occurs, and the core loads on the path, including hospitals, etc.

[0038] B2. Statistically analyze the number of affected devices, core load percentage, and predicted outage duration along the fault propagation path corresponding to the device in the target power grid. If the device's fault risk level in the target power grid is a potential fault, then if any one of the following conditions is met—core load percentage ≥ 8%, predicted outage duration ≥ 90 minutes, or number of affected devices ≥ 2—it is determined that preventive measures need to be initiated, and a proactive maintenance result is output. If all conditions are met simultaneously—core load percentage < 3%, predicted outage duration < 45 minutes, and number of affected devices = 1—it is determined that preventive measures do not need to be initiated, and a dynamically monitorable result is output. If the device's fault risk level in the target power grid is a low-incidence fault, then if any one of the following conditions is met—core load percentage ≥ 5%, predicted outage duration ≥ 30 minutes, and number of affected devices = 1 critical device—it is determined that preventive measures need to be initiated, and a proactive maintenance result is output. If all conditions are met simultaneously—core load percentage = 0%, predicted outage duration < 15 minutes, and number of affected devices = 1 non-critical device—it is determined that preventive measures do not need to be initiated, and a dynamically monitorable result is output.

[0039] It should be noted that, firstly, from the fault propagation path data output by the CAIO digital twin simulation, extract all device IDs covered by the path, including lines, transformers, circuit breakers, etc., and then, by comparing with the power grid equipment ledger, remove redundant devices without actual load association, and count the remaining devices to determine the number of affected devices. For core load percentage calculation, it is necessary to clearly define the power supply area covered by the fault propagation path. Filter the core loads within this area from the power grid load ledger, such as the total rated power of hospitals, data centers, and transportation hubs, and then calculate their percentage of the total rated power of the area's load. The formula is: Core Load Percentage = Total Core Load Power. The predicted outage duration is calculated as follows: (Rate / Total Area Load Power) × 100%. If the total area load data is missing, the load data collected in real time by CAIO can be used as a substitute. The predicted outage duration statistics need to be combined with the fault type and handling process. For emergency faults, the standard duration of "fault location (5 minutes) + reverse compensation (2 minutes) + power restoration (10 minutes)" is added together. For potential faults, the duration of "proactive maintenance preparation (60 minutes) + maintenance operation (30 minutes) + power restoration (15 minutes)" is added together. At the same time, the predicted outage duration is taken as the smaller value of the two values, referring to the "predicted outage duration when the fault is not intervened" in the digital twin simulation.

[0040] Step 3: Fault Location and Strategy Formulation: If a device in the target power grid needs to initiate preventive measures and requires proactive maintenance, the fault node of the device in the target power grid is located and analyzed to select the optimal EPA current disconnection point for the fault node of the device in the target power grid. At the same time, the reverse compensation current strategy corresponding to the optimal EPA of the fault node of the device in the target power grid is analyzed.

[0041] In a specific embodiment, the location analysis of the fault node of the device in the target power grid is performed as follows: First, the real-time data of 3-5 EPAs around the fault node are retrieved synchronously through P2P communication using an AI tracing model. The real-time data includes the fault current amplitude, fault direction, and voltage drop amplitude, and is standardized by unifying timestamps and removing outliers. The location is performed in four steps using a distributed boundary judgment algorithm: First, each EPA with the fault direction pointing to the same node is selected; then, the average fault current amplitude of each EPA with the same direction is calculated. If the average value is ≥1.5 times the line rated current, it is judged to be a short-circuit fault; then, based on the fault current traveling wave propagation time difference, the specific distance between the fault point and the nearest EPA is calculated according to "fault distance = wave speed × (EPA1 propagation time - EPA2 propagation time) / 2", thereby locating the specific tower number or equipment number corresponding to the device in the target power grid, and recording it as the fault node corresponding to the device in the target power grid.

[0042] In a specific embodiment, the process of selecting the optimal EPA current disconnection point for the fault node of the device in the target power grid is as follows: Prioritize selecting the EPA with the nearest specific tower number or device number in the target power grid. This EPA must also meet the following requirements: current load rate < 70%, historical circuit breaker opening and closing success rate ≥ 99.5%, and normal communication links with adjacent EPAs. If the nearest EPA does not meet these conditions, then filter sequentially from closest to furthest distance to select the optimal EPA current disconnection point for the fault node of the device in the target power grid.

[0043] In a specific embodiment, the analysis of the optimal EPA corresponding to the fault node of the device in the target power grid is carried out as follows: The fault type corresponding to the fault of the device in the target power grid is obtained. The fault types include single-phase ground fault, phase-to-phase short-circuit fault, and three-phase short-circuit fault. If the fault of the device in the target power grid is a single-phase ground fault, then the power grid-to-ground capacitance current calculation model is used: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. Where U is the rated line voltage of the faulted power grid, ω is the angular frequency, C is the line-to-ground capacitance per unit length of the fault node, and L is the length of the faulted line in the area. If the fault of the equipment in the target power grid is a phase-to-phase short-circuit fault, then according to the power grid-to-ground capacitance current calculation model: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. Where Z1 is the positive-sequence impedance of the fault node and Z2 is the negative-sequence impedance of the fault node, Z1=Z2 when the power grid is operating symmetrically. If the fault of the equipment in the target power grid is a three-phase short-circuit fault, then according to the power grid-to-ground capacitance current calculation model: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. After the optimal EPA of the fault node of the device in the target power grid is executed according to the corresponding reverse compensation current strategy, it is then analyzed whether the reverse compensation current strategy corresponding to the fault node of the device in the target power grid meets the standard.

[0044] In a specific embodiment, the analysis process for determining whether the reverse compensation current strategy corresponding to the fault node of the equipment in the target power grid meets the standard is as follows: After the optimal EPA of the fault node of the equipment in the target power grid completes the execution of the corresponding reverse compensation current strategy, CAIO retrieves the electrical quantity data of the fault node of the equipment and three surrounding EPAs in the target power grid. The three core indicators are: the fault node current of the fault node of the equipment and the three surrounding EPAs in the target power grid are all ≤0.5A, the voltage deviation of adjacent lines is all ≤±2%, and the temperature of the EPA compensation module is all <60℃. If all three indicators are met, it is determined that the reverse compensation current strategy corresponding to the fault node of the equipment in the target power grid meets the standard. If all three indicators are not met, it is determined that the reverse compensation current strategy corresponding to the fault node of the equipment in the target power grid does not meet the standard. At the same time, CAIO needs to trigger the power grid emergency response plan and immediately perform the current disconnection operation through the selected optimal EPA of the fault node.

[0045] Step 4: Precise Post-Fault Repair: After the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy, the repair plan corresponding to the fault node of the equipment in the target power grid is analyzed.

[0046] In a specific embodiment, the analysis of the maintenance plan corresponding to the fault node of the equipment in the target power grid is carried out as follows: The type and location of the faulty equipment in the target power grid are obtained, a spare parts list is generated, and the personnel closest to the fault point with the corresponding maintenance qualifications are matched based on the real-time location of the maintenance personnel's terminal. Maintenance tools are simultaneously dispatched to material stations near the fault node to ensure delivery within 30 minutes. At the same time, a safe working area is designated for the fault node of the equipment in the target power grid, and a three-dimensional operation guide and instructions for operation according to specifications are pushed to the maintenance personnel's terminal through an AR guidance module. Finally, the power restoration preparation strategy corresponding to the fault node of the equipment in the target power grid is analyzed.

[0047] Examples of embodiments of the present invention Figure 2 As shown, a power grid fault detection and maintenance system based on relay protection includes: a power grid data acquisition and networking module: used to acquire real-time data corresponding to each device in the target power grid through online monitoring devices and UAV inspection images; at the same time, deploying a collaborative autonomous network of CAIO and EPA on each device in the target power grid, and building a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO.

[0048] Fault Risk Assessment and Judgment Module: Based on the real-time data of each device in the target power grid, it assesses the fault risk level of each device in the target power grid at the current moment, and simulates the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, thereby determining whether each device in the target power grid needs to start preventive measures, and outputs the results of whether proactive maintenance or dynamic monitoring is required.

[0049] Fault location and strategy formulation module: When a device in the target power grid needs to take preventive measures and when it needs to perform proactive maintenance, the module performs fault node location analysis on the device in the target power grid, thereby selecting the optimal EPA current disconnection point for the device in the target power grid. At the same time, it analyzes the reverse compensation current strategy corresponding to the optimal EPA for the device in the target power grid.

[0050] Precise Post-Fault Repair Module: This module is used to analyze the repair plan corresponding to the fault node of the equipment in the target power grid after the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy.

[0051] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in this specification, they should all fall within the protection scope of the present invention.

Claims

1. A method for power grid fault detection and maintenance based on relay protection, characterized in that, include: Step 1: Power Grid Data Acquisition and Networking: Real-time data of each device in the target power grid is acquired through online monitoring devices and drone inspection images. At the same time, a collaborative autonomous network of CAIO and EPA is deployed on each device in the target power grid to build a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO. Step 2, Fault Risk Assessment and Determination: Based on the real-time data of each device in the target power grid, assess the fault risk level of each device in the target power grid at the current moment, and simulate the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, and then determine whether each device in the target power grid needs to start preventive measures, and output the results of whether proactive maintenance or dynamic monitoring is required. Step 3: Fault Location and Strategy Formulation: If a device in the target power grid needs to initiate preventive measures and is required to perform proactive maintenance, the fault node of the device in the target power grid is located and analyzed to select the optimal EPA current disconnection point for the fault node of the device in the target power grid. At the same time, the reverse compensation current strategy corresponding to the optimal EPA of the fault node of the device in the target power grid is analyzed. Step 4: Precise Post-Fault Repair: After the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy, the repair plan corresponding to the fault node of the equipment in the target power grid is analyzed.

2. The power grid fault detection and maintenance method based on relay protection as described in claim 1, characterized in that, The deployment process of the CAIO and EPA collaborative autonomous network in each device of the target power grid is as follows: CAIO is deployed on the cloud servers of each device in the target power grid, configured with a global data storage module and an ultra-real-time simulation engine, and connected to the power grid dispatch system to obtain topology and load basic data. EPA is embedded in each device and each smart circuit breaker in the target power grid. During deployment, according to the "regional division" principle, each device in the target power grid is divided into several EPA groups. Each group contains 3-5 adjacent EPAs. Each group establishes a dedicated C / S communication link with CAIO, and EPAs within the same group build a P2P communication subnet.

3. The power grid fault detection and maintenance method based on relay protection as described in claim 2, characterized in that, The specific assessment process for evaluating the equipment failure risk level of each device in the target power grid at the current moment is as follows: A1. Based on the real-time data of each device in the target power grid, obtain the three evaluation dimensions of each device in the target power grid. The three evaluation dimensions include electrical quantity dimension, equipment status dimension and environmental dimension. Then evaluate whether the electrical quantity dimension, equipment status dimension and environmental dimension of each device in the target power grid meet the qualification conditions. If any one of the three evaluation dimensions of a certain device in the target power grid does not meet the qualification conditions, then evaluate the equipment failure risk level of that device in the target power grid. A2. When assessing the equipment failure risk level of the corresponding device in the target power grid, the scores for each assessment dimension where the device fails to meet the qualification conditions are compared with the corresponding scores for the set assessment dimensions. The scores for each assessment dimension where the device fails to meet the qualification conditions are obtained. The scores for each assessment dimension where the device fails to meet the qualification conditions are accumulated and recorded as the total score for the device's assessment dimensions in the target power grid. If the total score for the device's assessment dimensions in the target power grid is greater than 20, it is recorded as a potential failure. If the total score for the device's assessment dimensions in the target power grid is less than or equal to 20, it is recorded as a low-incidence failure.

4. The power grid fault detection and maintenance method based on relay protection as described in claim 3, characterized in that, The specific simulation process for the fault impact range of each device in the simulated target power grid is as follows: B1. If a device in the target power grid triggers a potential fault or an emergency fault, the CAIO digital twin model is constructed based on the real-time topology and parameters of the target power grid. The operating data of the device in the target power grid is input, and the simulation duration is set. The simulation duration is 72 hours for potential faults and 1 hour for emergency faults. The fault propagation path corresponding to the device in the target power grid is output. B2. Statistically analyze the number of affected devices, core load percentage, and predicted outage duration along the fault propagation path corresponding to the device in the target power grid. If the device's fault risk level in the target power grid is a potential fault, then if any one of the following conditions is met—core load percentage ≥ 8%, predicted outage duration ≥ 90 minutes, or number of affected devices ≥ 2—it is determined that preventive measures need to be initiated, and a proactive maintenance result is output. If all conditions are met simultaneously—core load percentage < 3%, predicted outage duration < 45 minutes, and number of affected devices = 1—it is determined that preventive measures do not need to be initiated, and a dynamically monitorable result is output. If the device's fault risk level in the target power grid is a low-incidence fault, then if any one of the following conditions is met—core load percentage ≥ 5%, predicted outage duration ≥ 30 minutes, and number of affected devices = 1 critical device—it is determined that preventive measures need to be initiated, and a proactive maintenance result is output. If all conditions are met simultaneously—core load percentage = 0%, predicted outage duration < 15 minutes, and number of affected devices = 1 non-critical device—it is determined that preventive measures do not need to be initiated, and a dynamically monitorable result is output.

5. The power grid fault detection and maintenance method based on relay protection as described in claim 4, characterized in that, The specific analysis process for locating the fault node of the equipment in the target power grid is as follows: First, using an AI-based source tracing model and P2P communication, real-time data from 3-5 EPAs surrounding the fault node are retrieved synchronously. This real-time data includes fault current amplitude, fault direction, and voltage drop amplitude, and is standardized by unifying timestamps and removing outliers. Then, a distributed boundary judgment algorithm is used to locate the fault in four steps: First, EPAs with fault directions pointing to the same node are selected; then, the average fault current amplitude of each EPA with the same direction is calculated. If the average value is ≥1.5 times the line rated current, it is judged to be a short-circuit fault; then, based on the fault current traveling wave propagation time difference, the specific distance between the fault point and the nearest EPA is calculated according to "fault distance = wave speed × (EPA1 propagation time - EPA2 propagation time) / 2", thereby locating the specific tower number or equipment number corresponding to the equipment in the target power grid, and recording it as the fault node corresponding to the equipment in the target power grid.

6. The power grid fault detection and maintenance method based on relay protection as described in claim 5, characterized in that, The optimal EPA current disconnection point for this device in the target power grid is selected from the fault nodes. The specific selection process is as follows: Prioritize selecting the EPA with the nearest specific tower number or equipment number corresponding to the device in the target power grid. At the same time, the EPA must meet the following conditions: current load rate <70%, historical circuit breaker opening and closing success rate ≥99.5%, and normal communication links with adjacent EPAs. If the nearest EPA does not meet the conditions, filter in order of distance from nearest to farthest to select the EPA current disconnection point for the fault node of the device in the target power grid.

7. A power grid fault detection and maintenance method based on relay protection as described in claim 6, characterized in that, The analysis process for determining the optimal reverse compensation current strategy for the faulty node of the device in the target power grid is as follows: Obtain the fault type corresponding to the equipment fault in the target power grid. Fault types include single-phase ground fault, phase-to-phase short circuit fault, and three-phase short circuit fault. If the equipment fault in the target power grid is a single-phase ground fault, then calculate the grid-to-ground capacitance current according to the power grid ground capacitance current calculation model: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. Where U is the rated line voltage of the faulted power grid, ω is the angular frequency, C is the line-to-ground capacitance per unit length of the fault node, and L is the length of the faulted line in the area. If the fault of the equipment in the target power grid is a phase-to-phase short-circuit fault, then according to the power grid-to-ground capacitance current calculation model: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. Where Z1 is the positive-sequence impedance of the fault node and Z2 is the negative-sequence impedance of the fault node, Z1=Z2 when the power grid is operating symmetrically. If the fault of the equipment in the target power grid is a three-phase short-circuit fault, then according to the power grid-to-ground capacitance current calculation model: The reverse compensation current corresponding to the optimal EPA of the fault node of the equipment in the target power grid is obtained. After the optimal EPA of the fault node of the device in the target power grid is executed according to the corresponding reverse compensation current strategy, it is then analyzed whether the reverse compensation current strategy corresponding to the fault node of the device in the target power grid meets the standard.

8. The power grid fault detection and maintenance method based on relay protection as described in claim 7, characterized in that, The analysis process is as follows: Whether the reverse compensation current strategy corresponding to the fault node of the equipment in the target power grid meets the standard is analyzed in detail below. Once the optimal EPA for the faulty node in the target power grid has completed its execution according to the corresponding reverse compensation current strategy, CAIO retrieves the electrical quantity data of the faulty node and three surrounding EPAs in the target power grid. Three core indicators are required: the faulty node current of the faulty node and the three surrounding EPAs in the target power grid are all ≤0.5A, the voltage deviation of adjacent lines is all ≤±2%, and the temperature of the EPA compensation module is all <60℃. If all three indicators are met, the reverse compensation current strategy corresponding to the faulty node in the target power grid is deemed to be compliant. If all three indicators are not met, the reverse compensation current strategy corresponding to the faulty node in the target power grid is deemed to be non-compliant. At the same time, CAIO needs to trigger the power grid emergency response plan and immediately execute the current disconnection operation through the selected optimal EPA for the faulty node.

9. A power grid fault detection and maintenance method based on relay protection as described in claim 8, characterized in that, The analysis process for the maintenance plan corresponding to the faulty node of the equipment in the target power grid is as follows: The system obtains the faulty equipment type and location corresponding to the device in the target power grid, generates a spare parts list, and matches the nearest qualified personnel with the corresponding maintenance skills to the fault location using the real-time location of the maintenance personnel's terminal. Simultaneously, it dispatches maintenance tools to material stations near the fault node to ensure delivery within 30 minutes. At the same time, it delineates a safe working area for the faulty node of the device in the target power grid and pushes three-dimensional operation guidance and instructions for maintenance personnel to operate according to specifications to the maintenance personnel's terminal through the AR guidance module. It also analyzes the power restoration preparation strategy corresponding to the faulty node of the device in the target power grid.

10. A power grid fault detection and maintenance system based on relay protection, implementing the power grid fault detection and maintenance method based on relay protection as described in any one of claims 1-9, characterized in that, include: Power grid data acquisition and networking module: It is used to acquire real-time data of each device in the target power grid through online monitoring devices and drone inspection images. At the same time, it deploys a collaborative autonomous network of CAIO and EPA on each device in the target power grid, and builds a hybrid protocol network of P2P communication between EPA and C / S communication between device CAIO. Fault Risk Assessment and Judgment Module: Based on the real-time data of each device in the target power grid, it assesses the fault risk level of each device in the target power grid at the current moment, and simulates the fault impact range of each device in the target power grid by calling digital twin ultra-real-time simulation through CAIO, thereby determining whether each device in the target power grid needs to start preventive measures, and outputs the results of whether proactive maintenance or dynamic monitoring is required. Fault location and strategy formulation module: When a device in the target power grid needs to take preventive measures and when it needs to be proactively maintained, the module performs fault node location analysis on the device in the target power grid, thereby selecting the optimal EPA current disconnection point for the device in the target power grid. At the same time, it analyzes the reverse compensation current strategy corresponding to the optimal EPA for the device in the target power grid. Precise Post-Fault Repair Module: This module is used to analyze the repair plan corresponding to the fault node of the equipment in the target power grid after the optimal EPA of the fault node of the equipment in the target power grid forces the arc to be extinguished through the reverse compensation current strategy.

Images