Power distribution network fault early warning and isolation control method fused with current disturbance identification

Abstract

The invention discloses a power distribution network fault early warning and isolation control method fused with current disturbance identification, and relates to the technical field of power systems and intelligent control, and the method comprises the steps: S1, monitoring a current signal of a power distribution network in real time, and carrying out the current disturbance identification, S2, judging a fault type based on a current disturbance identification result, and S3, carrying out the fault early warning and isolation control based on the fault type. S4, according to the dynamically adjusted protection action constant value, executing fault isolation control; the power distribution network fault early warning and isolation control method fusing current disturbance identification realizes rapid linkage of fault early warning and isolation control, is suitable for various power distribution system architectures, and has good engineering application value.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power system and intelligent control technology, in particular to a power distribution network fault early warning and isolation control method fusing current disturbance identification. BACKGROUND

[0002] In the current operation of the power distribution network, the conventional fault early warning and isolation control method is mostly based on a static setting value strategy, that is, fixed current, voltage and other threshold parameters are set in advance to identify faults and trigger protection actions.

[0003] However, the operation environment of the power distribution network is complex and changeable, especially in the scenarios of large-scale distributed power access, dynamic changes in power grid topology and frequent load fluctuations. The fixed setting value strategy is difficult to fully adapt to diversified operating conditions. When a fault occurs, the protection device may malfunction, refuse to act or have insufficient protection coverage due to the failure to identify the fault type and its development characteristics in time, thereby reducing the operation reliability and fault handling efficiency of the system. In view of the above problems, researchers have begun to focus on fault identification mechanisms based on disturbance characteristics, especially the time domain and frequency domain characteristics of current disturbance under different fault conditions. Some existing research attempts to introduce disturbance signal analysis means to assist in identifying fault types, but lacks a deep integration control mechanism with the protection strategy, and has not yet formed a perfect dynamic setting value adjustment method. In addition, most of the existing methods have insufficient disturbance pattern identification capability, making it difficult to accurately match the fault scenarios and real-time operating conditions, resulting in delayed protection setting value updating and low accuracy. SUMMARY

[0004] The purpose of the present application is to provide a power distribution network fault early warning and isolation control method fusing current disturbance identification, which dynamically adjusts the protection action setting value according to the fault type to solve the problem that the fixed threshold value cannot adapt to complex operating conditions.

[0005] To achieve the above purpose, the present application provides the following technical scheme: a power distribution network fault early warning and isolation control method fusing current disturbance identification, the method comprising: S1, real-time monitoring of power distribution network current signals and current disturbance identification; S2, judging the fault type based on the current disturbance identification result; S3, dynamically adjusting the protection action setting value based on the fault type; S4, executing fault isolation control according to the dynamically adjusted protection action setting value.

[0006] Preferably, the current disturbance identification includes wavelet transform, Fourier transform or envelope analysis of the current signal to extract fault disturbance characteristic parameters, including disturbance amplitude, frequency distribution and transient starting point.

[0007] Preferably, the fault type judgment comprises feature matching on the extracted disturbance features, and classification by a trained fault identification model, the fault identification model being at least one of a support vector machine, a convolutional neural network or a long short-term memory network.

[0008] Preferably, the dynamic adjustment of the protection action setting value comprises correction of the setting parameters in the current operating state according to the fault type, the setting parameters including overcurrent protection setting value, instantaneous protection setting value, time delay protection time and directional protection setting value.

[0009] Preferably, the fault isolation control comprises control of switch device switching, feeder reconstruction and standby power supply access strategy, to ensure a minimum range of power outage area and the fastest power supply recovery.

[0010] Preferably, the process of dynamic adjustment of the protection action setting value supports real-time updating and online learning mechanism, for continuously optimizing the fault identification accuracy and protection action reliability.

[0011] Preferably, the parameter optimization of the dynamic adjustment of the protection action setting value adopts a multi-objective optimization algorithm for solution, the multi-objective optimization algorithm including particle swarm optimization, genetic algorithm or ant colony algorithm, to take into account the action speed, selectivity and sensitivity indicators.

[0012] Preferably, the fault types include single-phase grounding, two-phase short circuit, three-phase short circuit, intermittent grounding and arc grounding, different types of faults corresponding to different disturbance feature modes and protection strategies.

[0013] Preferably, the fault isolation control further comprises supply reconstruction path analysis on the non-fault area, to realize fast load transfer and non-fault area power supply recovery by using graph theory method or power distribution topology reconstruction algorithm.

[0014] Preferably, the method is deployed in a power distribution master station, an edge node or an intelligent terminal, and has distributed architecture adaptability, to realize flexible deployment and linkage response of fault identification and isolation control in centralized and distributed systems.

[0015] From the above technical solution, the present application has the following beneficial effects: The power distribution network fault early warning and isolation control method fusing current disturbance identification improves the accuracy and response speed of fault identification under complex working conditions, effectively avoids the problems of misoperation or refusal to operate caused by unreasonable fixed threshold setting, improves the sensitivity and reliability of protection operation, and enhances the processing capacity and overall safety level of the power distribution system for atypical faults, reduces the central control load, realizes the rapid linkage of fault early warning and isolation control, is suitable for various power distribution system architectures, and has good engineering application value. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The method flowchart of the present application is shown. DETAILED DESCRIPTION

[0017] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of protection of the present application.

[0018] As shown in the accompanying drawings, Figure 1 The present application provides a technical solution: a power distribution network fault early warning and isolation control method fusing current disturbance identification, which comprises: S1, real-time monitoring of power distribution network current signals and current disturbance identification; S2, determining the fault type based on the current disturbance identification result; S3, dynamically adjusting the protection action setting value based on the fault type; S4, performing fault isolation control according to the dynamically adjusted protection action setting value.

[0019] The method takes the current signal in the power distribution network system as the core monitoring object, and obtains three-phase current data containing transient and steady-state information through the current transformers and high-speed data acquisition modules deployed at each node. The system uses an embedded disturbance identification model to analyze the collected current signals in real time, uses multi-scale wavelet transform to decompose the current signals in the frequency domain, extracts transient feature quantities such as high-frequency energy change rate, zero-crossing point offset characteristics, and short-time amplitude mutation degree to identify potential disturbance events. On the basis of disturbance identification, by constructing a multi-dimensional feature vector and performing similarity matching with the pre-set templates in the fault feature library, the system can intelligently identify common power distribution network fault types, including single-phase ground fault, two-phase short circuit, three-phase short circuit, and high-impedance ground fault. After fault identification is completed, the system dynamically adjusts the protection strategy according to the fault nature and severity, and uses a fuzzy logic control algorithm to adjust the starting current value, time delay parameters, and protection sensitivity level of the over-current protection in real time, thereby improving the adaptability and selectivity of the protection system. The dynamic adjustment mechanism combines the current load state and historical operation data to ensure that the protection device has the best response characteristics under different operating scenarios. Finally, the system sends isolation instructions to the corresponding switch control unit through the communication module to control the switch equipment in the fault section to be quickly disconnected, realizes physical isolation of the fault source, prevents the fault from spreading to adjacent power supply areas, and at the same time preserves normal power supply in non-fault areas, thereby enhancing the power supply continuity and fault bearing capacity of the power distribution network.

[0020] The method can significantly improve the accuracy and response speed of power distribution network fault identification, effectively identify the difference between slight disturbance and serious fault through the fusion of current disturbance identification mechanism, and avoid misoperation and refusal to operate. Dynamic adjustment of protection action value makes the protection strategy more adaptive, improves the adaptive protection capability of the system, helps to quickly locate and remove the fault area, shortens the fault recovery time, and improves the power supply reliability of the power distribution network. In addition, automatic fault isolation control reduces manual intervention, improves the operation efficiency and the overall intelligent level of the system.

[0021] The current disturbance identification includes wavelet transform, Fourier transform or envelope analysis on the current signal to extract fault disturbance feature parameters, including disturbance amplitude, frequency distribution and transient starting point.

[0022] To achieve high-precision current disturbance identification, the embodiment performs multi-dimensional signal processing operations on the current signal after it is obtained to capture weak but critical disturbance characteristics. Specifically, the system uses wavelet transform to perform multi-scale decomposition on the time-domain signal, effectively extracts the transient starting point information at the initial stage of the fault, and identifies the current mutation point and its corresponding time position by analyzing the energy change trend of different scale coefficients. At the same time, the fast Fourier transform (FFT) technology is applied to analyze the frequency domain of the current signal, identify the main frequency component and its power distribution in the disturbance, and judge the harmonic disturbance or periodic fault characteristics. The envelope analysis method is further used to extract the envelope signal of non-stationary components and analyze its envelope energy density, envelope average amplitude and other characteristic indexes, so as to identify the weak disturbance signal caused by mechanical vibration, intermittent contact or high resistance fault. In the feature parameter extraction process, the system synchronously outputs the disturbance amplitude, frequency distribution and transient starting point and other key parameters, and inputs them into the fault identification module to support subsequent fault type judgment and protection strategy adjustment. The above algorithms are embedded in the edge computing node or the main control unit to ensure that the analysis result has a millisecond-level delay response, meeting the needs of real-time early warning and control.

[0023] By introducing various signal processing methods such as wavelet transform, Fourier transform and envelope analysis, this method can accurately identify disturbances with different frequency characteristics, different durations and different starting forms, and realize multi-dimensional modeling and comprehensive extraction of current disturbance characteristics. Especially in the face of complex fault scenarios (such as transient arc, high-frequency harmonic injection, etc.), this combined identification method has higher accuracy and robustness. At the same time, the extraction of disturbance amplitude, frequency distribution and transient starting point and other parameters provides high-quality data support for subsequent fault type judgment and protection setting optimization, thereby significantly improving the intelligence, accuracy and efficiency of the overall fault response system.

[0024] The fault type judgment includes feature matching on the extracted disturbance characteristics, and classification by a trained fault identification model, the fault identification model being at least one of a support vector machine, a convolutional neural network or a long short-term memory network.

[0025] To realize accurate judgment of fault type, after completing disturbance feature parameter extraction, the embodiment first constructs a standardized feature vector set, and the vector content includes multiple dimensions such as disturbance amplitude, frequency distribution, transient starting point, envelope energy, and spectrum form. The system matches the current disturbance feature vector with the preset typical fault template library through a feature matching mechanism to match the Euclidean distance or cosine similarity, and preliminarily identifies the type set to which the fault may belong. On this basis, a fault identification model is used for accurate classification. The model is trained by historical fault sample data, and can use a support vector machine (SVM) to divide the disturbance feature vector by a hyperplane to realize linear or nonlinear discrimination in a high-dimensional space; a convolutional neural network (CNN) can be used to perform deep convolution extraction on the two-dimensional spectrum graph of the disturbance signal to capture spatial local correlation features; and a long short-term memory network (LSTM) can be used to process the time dependence relationship in the disturbance signal sequence to enhance the modeling ability for complex nonlinear disturbance evolution process. The above model improves the classification accuracy through cross-validation and hyperparameter optimization in the training stage, and is deployed in an edge computing manner in the running stage to realize online identification of new disturbance data. Finally, the system feeds back the classification result to the protection strategy module to drive the corresponding fault response mechanism.

[0026] By introducing a machine learning model to intelligently identify fault types, compared with traditional rule-based classification methods, this method can automatically extract hidden patterns in disturbance features, significantly improving recognition accuracy and robustness. Support vector machines are suitable for small sample, high-dimensional feature space fault patterns; convolutional neural networks can process complex morphological, strongly coupled feature disturbance spectrum data; and long short-term memory networks have outstanding performance in processing nonlinear, time series faults, thereby enabling the system to have universal identification capability in multiple fault situations. In addition, the trainability and online updating mechanism of the model enable the fault identification system to have the potential for continuous optimization and self-evolution, which helps to improve the intelligent level and long-term adaptability of the power distribution network protection system.

[0027] The dynamic adjustment of the protection action setting value includes correcting the setting parameters in the current operating state according to the fault type, and the setting parameters include overcurrent protection setting value, instantaneous protection setting value, time delay protection time, and directional protection setting value.

[0028] After the fault type is identified, the system adjusts the setting parameters of the protection device in real time to match the fault nature and the current distribution network operating state to achieve a more accurate and dynamic response protection strategy. Specifically, the system first determines the adjustment strategy template based on the identified fault type (such as single-phase grounding, phase-to-phase short circuit or asymmetric fault), and extracts the corresponding strategy from the pre-set parameter correction database. The adjustment process includes four key protection settings: first, the overcurrent protection setting is dynamically increased or decreased according to the fault point current amplitude to ensure reliable action in low current fault or high resistance grounding scenarios; second, the instantaneous protection setting is corrected according to the fault location and real-time changes in system impedance, enabling it to quickly remove heavy faults; third, the time delay protection time adjusts the delay start time by combining fault current duration and load curve to prevent misoperation and improve selectivity; fourth, the directional protection setting is optimized according to the fault current direction and voltage phase angle relationship to enhance the ability to distinguish faults at the end of the feeder and the power supply side. All setting parameter corrections are implemented in a closed-loop control architecture, combining real-time operating data provided by the SCADA system and PMU (synchronous phasor measurement device) to achieve online adaptive configuration of the protection strategy.

[0029] Through dynamic adjustment of the setting parameters, this method significantly improves the sensitivity, selectivity and adaptability of the protection system. Accurate setting of overcurrent and instantaneous protection settings helps to accurately identify different levels of fault strength and achieve fast and effective fault removal; intelligent adjustment of the time delay protection time reduces the risk of misoperation while optimizing the coordination relationship of multi-level protection; adaptive updating of the directional protection setting enhances the ability to respond to multi-source power supply faults in complex feeder structures. In addition, the dynamic correction mechanism of the setting parameters avoids the lag of manual periodic adjustment, improves the overall response speed of the system and the timeliness of protection configuration, meeting the needs of the changing operating environment of modern distribution networks.

[0030] Fault isolation control includes controlling the switching equipment to jump together, feeder reconstruction and standby power supply access strategy to ensure the smallest range of power outage area and the fastest power supply recovery.

[0031] After the fault identification and setting parameter adjustment are completed, the embodiment enters a fault isolation control phase, a core target of which is to achieve accurate removal of a fault section and rapid recovery of power supply in a non-fault area. Specifically, the system first sends trip instructions to switch devices at the front and back of the fault point based on a real-time network topology graph and a fault positioning result, so as to quickly isolate the fault section. The switch devices can be automatic reclosing switches, vacuum circuit breakers or intelligent feeder terminals, and have remote control and local execution capabilities. Subsequently, the system performs topology reconstruction calculation on the current power distribution network structure, formulates an optimal feeder reconstruction scheme according to network connectivity, residual capacity and load priority. In the reconstruction process, the system selects a standby line or a non-fault branch, controls the corresponding switch to close, so that the load points not affected by the fault can obtain power supply again. In addition, the system also dispatches standby power sources, including diesel generators, energy storage devices or upper-level power sources, according to the current power supply capacity, and supplies power to the load nodes with higher priority through dynamic power matching and load distribution strategies. In the whole isolation and recovery process, the control strategy takes minimizing the number of power-off users and power supply interruption time as the target, fully utilizes multi-source power resources and network topology redundancy, and realizes rapid recovery and stable operation of the power distribution system.

[0032] The method significantly improves the response capability and power supply recovery efficiency of the power distribution system. Through automatic switch trip and close control, the fault section can be quickly removed to prevent the fault from expanding; the feeder reconstruction mechanism realizes flexible adjustment of the power supply path, effectively restores power supply in the non-fault area, and reduces the power-off range; and the standby power source access strategy improves the power supply guarantee capability of the system in the single power source failure scenario. The overall scheme realizes an adaptive and autonomous decision-making fault processing process, reduces the manual intervention time, reduces the operation and maintenance cost, improves the user power continuity and system power supply reliability. The system also has good expansibility and compatibility, can adapt to various power distribution network structures and operation modes, and provides key support for smart grid construction.

[0033] The process of dynamically adjusting the protection action setting value supports real-time updating and online learning mechanism, which is used to continuously optimize the fault identification accuracy and protection action reliability.

[0034] To realize the continuous optimization and adaptive enhancement of protection strategy, the embodiment further introduces real-time updating and online learning mechanism on the basis of dynamically adjusting protection action setting value, and constructs a data-driven dynamic optimization system. Specifically, the system collects and analyzes key data in the actual fault response process after fault identification and isolation control is completed, including disturbance signal waveform, fault positioning deviation, action time delay, protection misoperation rate and other indicators, to form a feedback data set. The data set is input to the online learning module, which updates the existing fault identification model and setting parameter optimization model based on incremental learning strategy or sliding window training mechanism. For example, the system can use online support vector regression (Online SVR) to real-time fit the relationship between fault current characteristics and setting value, or use incremental neural network to identify the pattern of protection action history to update the discrimination threshold and response weight. Under the condition of real-time data input, the model parameters are continuously adjusted, and the setting strategy evolves synchronously with the change of running state, load level and environment, so as to realize the dynamic optimization of protection action setting value. To ensure the stability and safety of the updating process, the system also has a strategy verification module, which uses a sandbox environment to simulate and verify the new strategy, and confirms that it does not affect the normal operation of the system before it can be deployed to the actual control link.

[0035] The embodiment significantly improves the intelligent level and dynamic adaptability of the protection system by introducing real-time updating and online learning mechanism. Compared with the traditional static setting method, this method can effectively deal with the protection configuration lag problem caused by frequent changes in the operation state of the distribution network, and realize more targeted and timely protection strategy adjustment. The online learning mechanism ensures that the fault identification model and setting strategy evolve over time and have long-term self-optimization ability, thereby improving the identification accuracy and reliability of protection action, reducing the occurrence rate of misoperation and refusal to operate, and enhancing the stability and power supply safety of the system.

[0036] The parameter optimization of dynamically adjusting protection action setting value is solved by using a multi-objective optimization algorithm, including particle swarm optimization, genetic algorithm or ant colony algorithm, to balance the action speed, selectivity and sensitivity indicators.

[0037] The embodiment introduces a multi-objective optimization strategy in the process of dynamically setting the protection action value, to achieve the balance and optimal configuration of the protection system among multiple performance indicators. First, the system determines the protection parameter set to be adjusted according to the fault identification result, including overcurrent setting value, delay setting value, instantaneous trip threshold and directional sensitivity coefficient, etc. Then, the system constructs a multi-objective optimization model, taking the evaluation indexes of action speed (i.e. minimizing the fault clearance response time), selectivity (avoiding false tripping of non-fault section) and sensitivity (reliable identification of slight faults) as the optimization objective function, and sets the mathematical expression and constraint conditions. In the optimization solving process, the particle swarm optimization (PSO) algorithm can be selected to search for the optimal parameter combination through the speed and position update rules of each particle in the group; the genetic algorithm (GA) can be used to evolve optimal solutions in the parameter space through selection, crossover and mutation genetic operations; or the ant colony optimization (ACO) algorithm can be used to simulate the pheromone mechanism to realize path optimization parameter configuration search. In each optimization iteration, the system evaluates the comprehensive performance of the current parameter configuration under the three objectives based on historical fault data and simulation results, and selects the optimal solution through the Pareto front analysis. After optimization, the parameter results are synchronized to the protection device in real time to realize adaptive adjustment.

[0038] The embodiment configures the protection action setting value through a multi-objective optimization algorithm, which not only overcomes the problem of ignoring multi-performance trade-off in traditional single-objective optimization methods, but also significantly improves the overall coordination and decision intelligence of the distribution protection system. Particle swarm optimization, genetic algorithm and ant colony optimization all have the advantages of strong global search ability and fast convergence speed, which can quickly find the optimal balance point considering action speed, protection selectivity and sensitivity in complex parameter space, so as to realize the configuration of protection behavior with faster fault clearance response, more accurate fault range and lower misoperation rate, and significantly enhance the practicality and robustness of the system.

[0039] The fault types include single-phase grounding, two-phase short circuit, three-phase short circuit, intermittent grounding and arc grounding, and different types of faults correspond to different disturbance characteristic patterns and protection strategies.

[0040] The embodiment realizes accurate identification and differentiated response to various typical fault types of power distribution network by constructing a rich fault feature library and a strategy mapping table. In the current disturbance identification stage, the system extracts key parameters including disturbance amplitude, frequency spectrum distribution, transient starting point, envelope fluctuation, etc. from the real-time collected current signal. For single-phase ground fault, the system usually identifies the characteristics of zero-sequence current mutation, fault-phase current rise, and non-fault-phase voltage rise, etc.; for two-phase short circuit, it shows two-phase current surge and voltage drop, and specific frequency harmonic components appear in the frequency spectrum; three-phase short circuit usually presents three-phase current symmetry sudden rise, and the energy of high-frequency components in the frequency spectrum is significantly enhanced. Intermittent ground fault presents periodic intermittent changes in current, discontinuous disturbance envelope, and fuzzy starting point; arc ground fault is usually accompanied by high-frequency noise signal and unstable zero-sequence current change, and its frequency spectrum shows wideband high-frequency distribution. The system matches and judges these disturbance patterns with the built-in fault type identification model, and accordingly calls the preset protection strategy: for example, for single-phase ground fault, delay low-sensitivity protection is adopted, for three-phase short circuit, high-speed trip protection is enabled, and for arc ground fault, arc extinguishing device linkage is guided. The whole process runs in real time in the main controller or distributed protection device, ensuring fast and effective switching to the most suitable protection action configuration under different fault scenarios.

[0041] By clearly distinguishing and modeling five typical fault types of power distribution network, the embodiment significantly improves the accuracy of fault identification and the pertinence of protection strategy. Each fault type is combined with its disturbance characteristics to establish a unique set of identification parameters and protection response mechanism, avoiding the misoperation and refusal of traditional general protection strategy under complex fault conditions. For example, arc ground fault is difficult to be identified by traditional low-frequency protection due to its high-frequency characteristics, but the present scheme can accurately identify and quickly handle it through exclusive frequency spectrum mode. The classification and processing of different fault types also help to optimize the allocation of fault handling resources, improve power supply reliability and system stability, and meet the intelligent protection needs under current complex power distribution environment.

[0042] Fault isolation control also includes supply reconstruction path analysis of non-fault area, which realizes fast load transfer and non-fault area power supply recovery by using graph theory method or power distribution topology reconstruction algorithm.

[0043] In this embodiment, when a fault occurs and the fault section isolation is completed, the system automatically starts the power supply reconstruction module to ensure uninterrupted power supply in the non-fault area. This module first constructs a complete network topology graph based on the real-time operating state of the distribution network, abstracts each node (such as a substation, feeder end, and load node) as a vertex of the graph, abstracts lines and switching devices as edges of the graph, and assigns attributes such as current capacity, state, direction, etc. After the fault isolation is completed, the system analyzes the feasible power supply paths in the current network through graph theory methods such as shortest path search, maximum flow calculation, and minimum cut set identification. If there are multiple candidate paths, the system further introduces weight indicators such as power supply capacity, power supply distance, and load level, and uses a weighted optimal path algorithm for reconstruction path selection. At the same time, distribution topology reconstruction algorithms such as a mixed integer linear programming (MILP) based distribution network power flow optimization model or heuristic search strategies (such as tabu search and ant colony optimization) can be applied to dynamically reconstruct the entire network topology, allowing the load to quickly connect from the standby or non-fault feeder. After the path selection is completed, the system issues a closing instruction to the corresponding switching device, completes the physical path switching, and realizes load transfer and rapid power supply recovery in the non-fault area.

[0044] This method effectively improves the load reconfiguration capability of the distribution network under fault conditions through power supply reconstruction path analysis, significantly shortening the power outage time in the non-fault area. The introduction of graph theory and topology reconstruction algorithms not only enhances the global optimality of path selection, but also improves the automation level and computational efficiency of the system in dealing with complex power supply network structures. The rapid load transfer mechanism realizes the priority protection of important user power supply, reduces the dependence on standby power supply, and improves the utilization rate of main network power supply resources. Through automatic path identification and control, this method reduces the time of manual intervention, improves the operation and maintenance efficiency and emergency response capability, and provides strong support for the robustness and flexibility of modern intelligent distribution systems.

[0045] The method is deployed in a distribution main station, edge node, or intelligent terminal, has distributed architecture adaptation capability, and realizes flexible deployment and linkage response of fault identification and isolation control in centralized and distributed systems.

[0046] To meet the demand of fault response capability of power distribution systems with different scales and structures, the embodiment is designed as a flexible structure deployable on multi-level hardware platforms, supporting running in power distribution master stations, edge nodes and intelligent terminals. The system overall adopts a distributed architecture design, and core function modules such as current disturbance identification, fault classification judgment, protection action setting adjustment and fault isolation control can be distributed and deployed as needed. For example, in a centralized architecture, all functions are deployed in the power distribution master station, which is suitable for large urban power grids with strong unified control capability; in a distributed or hybrid architecture, edge nodes (such as substation controllers) and intelligent terminals (such as feeder terminals FTU and power distribution automation devices) respectively undertake local data acquisition, rapid disturbance identification and primary fault isolation control tasks, while the master station is responsible for global decision-making and coordination. The system uses a unified communication protocol (such as IEC60870-5-104, IEC 61850) to realize multi-level cooperation, and the edge nodes can independently execute local protection strategies when the master station communication is interrupted, improving the fault tolerance and autonomy of the system. At the same time, each function unit has dynamic registration and resource scheduling capability, supporting task migration or parallel execution between devices, meeting the high reliability and high flexibility operation requirements.

[0047] The deployment architecture gives the system great flexibility and scalability, enabling it to adapt to power distribution networks of various scales and structures. Through distributed deployment, fault identification and isolation response can be completed locally quickly, greatly improving response speed; centralized coordination can achieve optimal protection strategy formulation and reconstruction path scheduling for the entire network, taking into account global and local control requirements. The multi-level linkage capability and modular design of the system not only improve the efficiency and accuracy of fault handling, but also enhance the stable operation capability in extreme situations such as communication anomalies or hardware failures. In addition, the architecture provides a good foundation for future access to distributed energy, microgrids and new power electronic devices, etc., with good compatibility and continuous evolution potential.

[0048] Although embodiments of the present application have been shown and described, it is to be understood that various modifications, substitutions, replacements and changes can be made to these embodiments without departing from the principles and spirit of the present application, and the scope of the present application is defined by the appended claims and their equivalents.

Claims

1. A distribution network fault warning and isolation control method integrating current disturbance identification, characterized in that: The method comprises: S1. Real-time monitoring of distribution network current signals and identification of current disturbances; S2. Determine the fault type based on the current disturbance identification result; S3. Dynamically adjust the protection action setting based on the fault type; S4. Execute fault isolation control according to the dynamically adjusted protection action setting.

2. A distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1, characterized in that: The current disturbance identification includes performing wavelet transform, Fourier transform or envelope analysis on the current signal to extract fault disturbance characteristic parameters, and the fault disturbance characteristic parameters include disturbance amplitude, frequency distribution and transient starting point.

3. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The fault type judgment includes feature matching of the extracted disturbance features and classification through a fault recognition model obtained through training. The fault recognition model is at least one of a support vector machine, a convolutional neural network or a long short-term memory network.

4. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The dynamic adjustment of the protection action setting value includes modifying the setting parameters under the current operating state according to the fault type, and the setting parameters include the overcurrent protection setting value, the quick-break protection setting value, the delayed protection time and the directional protection setting value.

5. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The fault isolation control includes controlling the tripping and closing of switchgear, feeder reconstruction, and backup power access strategy to ensure the smallest power outage area and the fastest power restoration.

6. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The process of dynamically adjusting the protection action setting supports real-time updating and online learning mechanisms for continuously optimizing the fault identification accuracy and protection action reliability.

7. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The parameter optimization of the dynamic adjustment of the protection action constant value is solved by using a multi-objective optimization algorithm, which includes a particle swarm optimization, a genetic algorithm or an ant colony algorithm to take into account the action speed, selectivity and sensitivity indicators.

8. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The fault types include single-phase grounding, two-phase short circuit, three-phase short circuit, intermittent grounding, and arc grounding. Different types of faults correspond to different disturbance characteristic modes and protection strategies.

9. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The fault isolation control also includes performing power supply reconstruction path analysis on non-fault areas, and utilizing graph theory methods or distribution topology reconstruction algorithms to achieve rapid load transfer and power supply restoration in non-fault areas.

10. The distribution network fault early warning and isolation control method integrating current disturbance identification according to claim 1 is characterized in that: The method is deployed in a power distribution master station, edge node or smart terminal, and has the ability to adapt to a distributed architecture to achieve flexible deployment and linkage response of fault identification and isolation control in centralized and distributed systems.

Images