Distribution line fault automatic positioning and isolation method

Abstract

The invention relates to the technical field of fault detection, in particular to a distribution line fault automatic positioning and isolating method, which comprises the following steps of: installing an intelligent fault indicator at each distribution node of a distribution network; the data of all the intelligent fault indicators are transmitted to the central processor in real time by using a wireless communication technology; the central processor predicts and positions potential line fault points by using a machine learning algorithm according to the received data; after a fault point is detected, an optimal isolation scheme is calculated, and rapid isolation of a fault area is realized by controlling an intelligent circuit breaker or an isolation switch in a power distribution network; after the fault is isolated, the dynamic self-adaptive control module is automatically started, and the dynamic self-adaptive control module dynamically adjusts the load distribution and the current path of each node in the power distribution network based on the real-time monitoring data and the network state. According to the method, the fault response capability of the power distribution network is remarkably improved, the fault processing time is shortened, and the loss and inconvenience caused by the fault are reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a method for automatically locating and isolating a distribution line fault. Background Art

[0002] In modern power systems, the stable and reliable operation of distribution networks is crucial to ensuring the normal operation of society. With the increasing complexity of power grids and the continuous growth of load demand, the rapid location and effective isolation of line faults have become a major challenge in the operation and maintenance of power systems. Traditional fault handling methods rely on manual detection and mechanical protection devices, which not only have a long response time, but also make it difficult to accurately and quickly locate the fault point, resulting in a prolonged fault recovery time. In severe cases, it may even affect the normal power supply of a large number of users.

[0003] In addition, traditional methods are inefficient in grid restoration operations after a fault occurs. Due to the lack of real-time monitoring and intelligent network management systems, grid reconstruction and load redistribution after fault isolation often take a long time to complete, which not only increases the scope and duration of power outages, but may also pose a further threat to the stability of the grid.

[0004] To address these issues, although some automated fault detection and location solutions have been proposed in the prior art, these methods often rely on dense sensor networks and complex communication facilities, increasing the construction and maintenance costs of the system. Summary of the invention

[0005] Based on the above purpose, the present invention provides a method for automatically locating and isolating distribution line faults.

[0006] The technical solution of the present invention is: a method for automatically locating and isolating a distribution line fault, comprising the following steps: S1: Install intelligent fault indicators at each distribution node of the distribution network. Each intelligent fault indicator can independently monitor the current and voltage passing through, as well as the temperature and vibration of the line; S2: Using wireless communication technology, the data of all intelligent fault indicators are transmitted to the central processor in real time; S3: The central processor uses machine learning algorithms to predict and locate potential line fault points based on the received data; S4: After the fault point is detected, the optimal isolation solution is calculated, and the fault area is quickly isolated by controlling the intelligent circuit breaker or disconnector in the distribution network; S5: After the fault is isolated, the dynamic adaptive control module is automatically started. The dynamic adaptive control module dynamically adjusts the load distribution and current path of each node in the distribution network based on real-time monitoring data and network status.

[0007] Furthermore, the distribution nodes include substation entry and exit nodes, trunk line nodes, branch nodes, smart transformer nodes, and microgrid or distributed energy access points.

[0008] Furthermore, step S2 specifically includes: S21: Each intelligent fault indicator is equipped with a micro wireless transmission module that supports at least one wireless communication protocol, including Bluetooth, Wi-Fi, LoRaWAN, or NB-IoT, to achieve communication with the central processor; S22: The intelligent fault indicator regularly collects the current, voltage, temperature and vibration data of the node where it is located, and preliminarily analyzes the data to determine whether there is an abnormal situation; S23: When the intelligent fault indicator detects an abnormal situation or at its predetermined data collection time point, it will package the data and send it to the central processor through its wireless transmission module.

[0009] Furthermore, step S3 specifically includes: S31: The central processor collects and stores historical and real-time data from each intelligent fault indicator, including current, voltage, temperature and vibration information, to form a comprehensive data set; S32: Based on the comprehensive data set, the central processor uses data preprocessing techniques to optimize the input data quality of the machine learning model; S33: The central processor uses random forest for model training. The training process includes supervised learning using data from known fault events and cross-validation of the model. S34: After the training is completed, the central processor will input the real-time received data into the trained machine learning model, and the model will analyze the data to predict the fault type and location of the fault point.

[0010] Furthermore, the random forest model training specifically includes: Construct training and test sets: Divide the comprehensive dataset into two parts: one for training the random forest model and the other for testing and verifying the performance of the model; Use the random forest algorithm to learn the training set, improve the prediction accuracy by building multiple decision trees and combining the prediction results. During the training process, the model is optimized by adjusting the number of decision trees, the depth of the tree, and the parameters of the feature selection method; Use the test set to evaluate the accuracy and generalization ability of the random forest model, and tune the model by analyzing the difference between the model prediction results and the actual fault data; The data collected in real time is input into the trained random forest model, which will predict the failure probability of each data point. Based on the prediction results, the central processor identifies nodes or line segments with high failure probabilities, thereby quickly locating potential failure points.

[0011] Furthermore, the model tuning specifically includes: Regularly compare the fault results predicted by the model with the actual faults that occurred, and analyze the model's prediction accuracy, false positive rate, and false negative rate performance indicators; In the case of discrepancies between model predictions and actual faults, analyze the causes of the discrepancies, including improper feature selection, insufficient data preprocessing, and suboptimal model hyperparameter settings. Based on the results of the discrepancy analysis, re-evaluate and select data features that are more representative or better reflect fault characteristics. Adjust the model's hyperparameters, including the number and depth of decision trees, and retrain the model using updated features and data. Through continuous iterations of training, prediction, evaluation, and tuning, the model's prediction accuracy and generalization ability can be gradually improved.

[0012] Furthermore, step S4 specifically includes: S41: After receiving the fault signal from the intelligent fault indicator, the central processor first determines the location of the fault point and the fault type; S42: The central processor uses the topological information of the power distribution network to analyze and determine all isolation paths, taking into account the importance of power supply in the fault area, the stability of the network after the fault, and the number of affected users; S43: Based on the location of the fault point and all isolation paths, the central processor uses an optimization algorithm to calculate the best isolation solution, which aims to minimize the impact of the fault on the network and users, while ensuring rapid isolation of the fault area and maximum stability of power supply in non-fault areas.

[0013] Furthermore, after determining the best isolation scheme in step S4, the central processor issues instructions through wireless communication technology to control the corresponding intelligent circuit breakers or isolating switches in the distribution network to perform isolation operations, including remotely modifying the opening and closing states of the circuit breakers or switches to ensure that the fault area is quickly isolated. After the fault isolation operation is performed, the central processor will monitor the network load and power supply conditions outside the fault area, and adjust network operating parameters or start the backup power supply plan to maintain stable operation of the network.

[0014] Furthermore, step S5 specifically includes: S51: The dynamic adaptive control module collects real-time data from each intelligent fault indicator in the power distribution network, including current, voltage, frequency and temperature information, as well as network topology change information after fault isolation; S52: The collected data will be used to update the state model of the distribution network in real time, which reflects the current network load distribution, voltage level and possible power supply paths; S53: Based on the updated network state model, the dynamic adaptive control module uses the shortest path algorithm to calculate the optimal load distribution and current path. This calculation process aims to achieve two main goals: one is to minimize the impact of fault isolation on user power supply, and the other is to ensure the stability and safety of power grid operation; After determining the optimal load distribution and current path, the dynamic adaptive control module sends control instructions to the relevant intelligent circuit breakers, disconnectors and load management equipment through wireless communication technology, guiding the adjustment of operating status according to the optimization results, including adjusting the opening and closing of circuit breakers, changing the connection path of disconnectors or adjusting the load output of key nodes.

[0015] Beneficial effects of the present invention: The present invention installs intelligent fault indicators at key nodes of the distribution network and uses wireless communication technology to transmit monitoring data to the central processor in real time. The present invention can quickly and accurately predict and locate line fault points. The central processor uses advanced machine learning algorithms, which not only greatly improves the speed and accuracy of fault location, but also calculates the best isolation plan by analyzing the fault type and impact, and realizes rapid isolation of the fault area. This process significantly improves the distribution network's response capability to faults, reduces fault processing time, and reduces losses and inconveniences caused by faults.

[0016] The dynamic adaptive control module introduced in the present invention can dynamically adjust load distribution and current path based on real-time monitoring data and distribution network status. This intelligent adjustment process ensures that even after a fault occurs and is isolated, the grid operation status can be quickly reconfigured to optimize power distribution and minimize the impact on user power supply. In addition, this module can also optimize the fault recovery process, improve the flexibility and efficiency of distribution network recovery, and ensure grid stability and user satisfaction. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings in the following description are only for the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 The figure is a schematic diagram of a method flow of an embodiment of the present invention. DETAILED DESCRIPTION

[0019] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with specific embodiments.

[0020] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the present invention should be understood by people with ordinary skills in the field to which the present invention belongs. The "first", "second" and similar words used in the present invention do not indicate any order, quantity or importance, but are only used to distinguish different components. "Include" or "comprise" and similar words mean that the elements or objects appearing before the word include the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. "Up", "down", "left", "right" and the like are only used to indicate relative positional relationships. When the absolute position of the described object changes, the relative positional relationship may also change accordingly.

[0021] like Figure 1 As shown, the method for automatically locating and isolating a distribution line fault comprises the following steps: S1: Install intelligent fault indicators at each distribution node of the distribution network. Each intelligent fault indicator can independently monitor the current and voltage passing through, as well as the temperature and vibration of the line; S2: Using wireless communication technology, the data of all intelligent fault indicators are transmitted to the central processor in real time; S3: The central processor uses machine learning algorithms to predict and locate potential line fault points based on the received data; S4: After the fault point is detected, the optimal isolation solution is calculated, and the fault area is quickly isolated by controlling the intelligent circuit breaker or disconnector in the distribution network; S5: After the fault is isolated, the dynamic adaptive control module is automatically started. Based on real-time monitoring data and network status, the dynamic adaptive control module dynamically adjusts the load distribution and current path of each node in the distribution network, reduces the impact on user power supply, and improves the flexibility and efficiency of system recovery.

[0022] Distribution nodes include substation entry and exit nodes, trunk line nodes, branch nodes, smart transformer nodes, and microgrid or distributed energy access points.

[0023] S2 specifically includes: S21: Each intelligent fault indicator is equipped with a micro wireless transmission module that supports at least one wireless communication protocol, including Bluetooth, Wi-Fi, LoRaWAN, or NB-IoT, to achieve communication with the central processor; S22: The intelligent fault indicator regularly collects the current, voltage, temperature and vibration data of the node where it is located, and preliminarily analyzes the data to determine whether there is an abnormal situation; S23: When the intelligent fault indicator detects an abnormal situation or at its predetermined data collection time point, it will package the data and send it to the central processor through its wireless transmission module.

[0024] S3 specifically includes: S31: The central processor collects and stores historical and real-time data from each intelligent fault indicator, including current, voltage, temperature and vibration information, to form a comprehensive data set; S32: Based on the comprehensive data set, the central processor uses data preprocessing techniques to optimize the input data quality of the machine learning model; S33: The central processor uses random forest for model training. The training process includes supervised learning using data from known fault events and cross-validation of the model. S34: After the training is completed, the central processor inputs the real-time received data into the trained machine learning model, and the model analyzes the data to predict the fault type and the location of the fault point; When the machine learning model predicts a potential fault event, the central processor will further analyze the probability of the event, the possible scope of impact, and the optimal isolation strategy in order to take quick and effective response measures; the central processor will also continuously monitor the model's predictive performance and regularly retrain and optimize the model using the latest data sets to adapt to changes in distribution network conditions and improve the accuracy of fault prediction.

[0025] Random forest model training specifically includes: Construct training and test sets: Divide the comprehensive dataset into two parts: one for training the random forest model and the other for testing and verifying the performance of the model; Use the random forest algorithm to learn the training set, improve the prediction accuracy by building multiple decision trees and combining the prediction results. During the training process, the model is optimized by adjusting the number of decision trees, the depth of the tree, and the parameters of the feature selection method; Use the test set to evaluate the accuracy and generalization ability of the random forest model, and tune the model by analyzing the difference between the model prediction results and the actual fault data; The real-time collected data is input into the trained random forest model, which predicts the failure probability of each data point. Based on the prediction results, the central processor identifies nodes or line segments with high failure probability, thus quickly locating potential failure points. Over time, the random forest model is regularly retrained and updated with newly collected fault and non-fault data to adapt to changes in the distribution network operating environment, ensuring continuous optimization of fault prediction and location capabilities.

[0026] Model tuning specifically includes: Regularly compare the fault results predicted by the model with the actual faults that occurred, and analyze the model's prediction accuracy, false positive rate, and false negative rate performance indicators; In the case of discrepancies between model predictions and actual faults, analyze the causes of the discrepancies, including improper feature selection, insufficient data preprocessing, and suboptimal model hyperparameter settings. Based on the results of the discrepancy analysis, re-evaluate and select data features that are more representative or better reflect fault characteristics. Adjust the model's hyperparameters, including the number and depth of decision trees, and retrain the model using updated features and data. Through continuous iterations of training, prediction, evaluation, and tuning, the model's prediction accuracy and generalization ability can be gradually improved.

[0027] S4 specifically includes: S41: After receiving the fault signal from the intelligent fault indicator, the central processor first determines the location of the fault point and the fault type; S42: The central processor uses the topological information of the power distribution network to analyze and determine all isolation paths, taking into account the importance of power supply in the fault area, the stability of the network after the fault, and the number of affected users; S43: Based on the location of the fault point and all isolation paths, the central processor uses an optimization algorithm to calculate the best isolation solution, which aims to minimize the impact of the fault on the network and users, while ensuring rapid isolation of the fault area and maximum stability of power supply in the non-fault area; Among them, optimization algorithm is a computational method used to find the optimal or satisfactory solution to a problem. Such algorithms are widely used in many fields such as mathematics, computer science, engineering, economics, etc., and the range of problems they solve ranges from simple linear programming to complex nonlinear and multi-objective optimization problems; Depending on the nature of the problem and the solution method, suitable optimization algorithms include but are not limited to: Dijkstra algorithm: It is suitable for solving the single-source shortest path problem, that is, the shortest path from a given vertex to all other vertices in the graph. Although it is not specifically used to deal with global optimization problems such as TSP, it can be used as a basic tool for some simple cases.

[0028] A* search algorithm: It combines the advantages of heuristic methods and Dijkstra algorithm, and can find the optimal path more efficiently when the target location is known. It guides the search process by introducing a heuristic function to evaluate the estimated distance from the current node to the target.

[0029] Dynamic programming: It can be used to solve optimization tasks with overlapping subproblems, such as simple TSP variants. It is very effective for smaller problems, but the computational complexity increases rapidly as the problem size increases.

[0030] Genetic Algorithm (GA): An artificial intelligence technology that imitates the selection mechanism of nature. It simulates the natural evolution process by encoding individuals, selection, crossover, mutation and other operations to search for the global optimal solution. It is suitable for large-scale problems that are difficult to solve directly.

[0031] Particle Swarm Optimization (PSO): A swarm intelligence algorithm inspired by the foraging behavior of bird flocks, which searches for the global optimal solution by moving multiple particles in the solution space. It is suitable for optimization problems in continuous space.

[0032] Ant Colony Algorithm (ACO): It simulates the behavior of ants releasing pheromones in the process of searching for food, and gradually constructs one or more optimal paths. It is particularly suitable for combinatorial optimization problems such as TSP.

[0033] Branch and bound method: Reduce unnecessary exploration by pruning the search tree, thereby improving efficiency. Applicable to discrete optimization problems.

[0034] Local search algorithms: including hill climbing and taboo search, which continuously adjust the current solution to improve its quality until no further improvement is possible. These methods can often find good approximate solutions.

[0035] The optimization algorithm in the embodiment of the present invention may specifically be an A* search algorithm, which is an existing conventional technical solution, and the specific implementation steps are not described in detail here.

[0036] After the best isolation scheme is determined in S4, the central processor issues instructions through wireless communication technology to control the corresponding intelligent circuit breakers or disconnectors in the distribution network to perform isolation operations, including remotely modifying the opening and closing status of the circuit breakers or switches to ensure that the fault area is quickly isolated. After the fault isolation operation is performed, the central processor will monitor the network load and power supply conditions outside the fault area, and adjust the network operating parameters or start the backup power supply plan to maintain the stable operation of the network.

[0037] S5 specifically includes: S51: The dynamic adaptive control module collects real-time data from each intelligent fault indicator in the power distribution network, including current, voltage, frequency and temperature information, as well as network topology change information after fault isolation; S52: The collected data will be used to update the state model of the distribution network in real time, which reflects the current network load distribution, voltage level and possible power supply paths; S53: Based on the updated network state model, the dynamic adaptive control module uses the shortest path algorithm to calculate the optimal load distribution and current path. This calculation process aims to achieve two main goals: one is to minimize the impact of fault isolation on user power supply, and the other is to ensure the stability and safety of power grid operation; Among them, the shortest path algorithm is a basic problem in graph theory, which is used to find the shortest path from a starting point to other vertices in a weighted graph. According to different requirements and graph characteristics, there are several commonly used shortest path algorithms: Dijkstra Algorithm: Applicable scenarios: Applicable to graphs with non-negative edge weights.

[0038] Features: Guaranteed to find the single-source shortest path, that is, the shortest path from a given source point to all other vertices.

[0039] Time complexity: For dense graphs, it is usually O(V^2), which can be optimized to O((V+E) logV) using a priority queue.

[0040] Bellman-Ford Algorithm: Applicable scenarios: Applicable to situations where edge weights can be negative (but there cannot be negative weight cycles).

[0041] Features: Able to detect negative weight cycles and can handle graphs with negative weight edges.

[0042] Time complexity: O(VE), where V is the number of vertices and E is the number of edges.

[0043] Floyd-Warshall algorithm: Applicable scenarios: Applicable to solving the shortest path between all vertex pairs.

[0044] Features: Dynamic programming method, suitable for dense graphs.

[0045] Time complexity: O(V^3), suitable for situations where the number of vertices is not large.

[0046] A* search algorithm: Applicable scenarios: Especially suitable for graphs with heuristic information, such as map navigation.

[0047] Features: It combines the best-first search strategy of Dijkstra algorithm and the greedy best-first search strategy, and estimates the distance from the current node to the target through a heuristic function.

[0048] Time complexity: depends on the quality of the heuristic function, ideally close to linear time.

[0049] Johnson's algorithm: Applicable scenarios: Used to solve the shortest path problem between all vertex pairs in a sparse graph.

[0050] Features: First, reweight all edges so that the weights of all edges become non-negative, and then run the Dijkstra algorithm once for each vertex.

[0051] Time complexity: O(VE + V^2 log V).

[0052] SPFA (Shortest Path Faster Algorithm): Applicable scenarios: Similar to Bellman-Ford, but implemented through queues, which can be faster than Bellman-Ford in some cases.

[0053] Characteristics: May perform poorly when there are many negative-weight edges in the graph.

[0054] Time complexity: Good in the average case, but can reach O(VE) in the worst case.

[0055] The shortest path algorithm in the embodiment of the present invention may specifically be a Dijkstra algorithm, which is an existing conventional technical solution, and the specific implementation steps are not described in detail here; After determining the optimal load distribution and current path, the dynamic adaptive control module sends control instructions to the relevant intelligent circuit breakers, disconnectors and load management equipment through wireless communication technology, guiding them to adjust the operating status according to the optimization results, including adjusting the opening and closing of circuit breakers, changing the connection path of disconnectors or adjusting the load output of key nodes; While performing load and path adjustment operations, the dynamic adaptive control module continuously monitors network status and device feedback to ensure that all adjustment measures achieve the expected results. If an unexpected change or new fault condition is encountered, the module will restart the optimization process to quickly adapt to the new network status and ensure the continuity of power supply and network stability. In addition, the dynamic adaptive control module can also record and learn the data and experience obtained from each event processing, and continuously optimize its algorithm to improve the efficiency and accuracy of future event response.

[0056] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present invention is limited to these examples. Under the concept of the present invention, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the present invention as described above, which are not provided in detail for the sake of simplicity.

[0057] The present invention is intended to cover all such substitutions, modifications and variations that fall within the broad scope of the claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for automatically locating and isolating a distribution line fault, characterized in that: The following steps are involved: S1: Install intelligent fault indicators at each distribution node of the distribution network. Each intelligent fault indicator can independently monitor the current and voltage passing through, as well as the temperature and vibration of the line; S2: Using wireless communication technology, the data of all intelligent fault indicators are transmitted to the central processor in real time; S3: The central processor uses machine learning algorithms to predict and locate potential line fault points based on the received data; S4: After the fault point is detected, the optimal isolation solution is calculated, and the fault area is quickly isolated by controlling the intelligent circuit breaker or disconnector in the distribution network; S5: After the fault is isolated, the dynamic adaptive control module is automatically started. The dynamic adaptive control module dynamically adjusts the load distribution and current path of each node in the distribution network based on real-time monitoring data and network status.

2. The method for automatically locating and isolating a distribution line fault according to claim 1, characterized in that: The distribution nodes include substation entry and exit nodes, trunk line nodes, branch nodes, smart transformer nodes, and microgrid or distributed energy access points.

3. The method for automatically locating and isolating a distribution line fault according to claim 2, characterized in that: Step S2 specifically includes: S21: Each intelligent fault indicator is equipped with a micro wireless transmission module that supports at least one wireless communication protocol, including Bluetooth, Wi-Fi, LoRaWAN, or NB-IoT, to achieve communication with the central processor; S22: The intelligent fault indicator regularly collects the current, voltage, temperature and vibration data of the node where it is located, and preliminarily analyzes the data to determine whether there is an abnormal situation; S23: When the intelligent fault indicator detects an abnormal situation or at its predetermined data collection time point, it will package the data and send it to the central processor through its wireless transmission module.

4. The method for automatically locating and isolating a distribution line fault according to claim 3, characterized in that: Step S3 specifically includes: S31: The central processor collects and stores historical and real-time data from each intelligent fault indicator, including current, voltage, temperature and vibration information, to form a comprehensive data set; S32: Based on the comprehensive data set, the central processor uses data preprocessing techniques to optimize the input data quality of the machine learning model; S33: The central processor uses random forest for model training. The training process includes supervised learning using data from known fault events and cross-validation of the model. S34: After the training is completed, the central processor will input the real-time received data into the trained machine learning model, and the model will analyze the data to predict the fault type and location of the fault point.

5. The method for automatically locating and isolating a distribution line fault according to claim 4, characterized in that: The random forest model training specifically includes: Construct training and test sets: Divide the comprehensive dataset into two parts: one for training the random forest model and the other for testing and verifying the performance of the model; Use the random forest algorithm to learn the training set, improve the prediction accuracy by building multiple decision trees and combining the prediction results. During the training process, the model is optimized by adjusting the number of decision trees, the depth of the tree, and the parameters of the feature selection method; Use the test set to evaluate the accuracy and generalization ability of the random forest model, and tune the model by analyzing the difference between the model prediction results and the actual fault data; The data collected in real time is input into the trained random forest model, which will predict the failure probability of each data point. Based on the prediction results, the central processor identifies nodes or line segments with high failure probabilities, thereby quickly locating potential failure points.

6. The method for automatically locating and isolating a distribution line fault according to claim 5, characterized in that: The model tuning specifically includes: Regularly compare the fault results predicted by the model with the actual faults that occurred, and analyze the model's prediction accuracy, false positive rate, and false negative rate performance indicators; In the case of discrepancies between model predictions and actual faults, analyze the causes of the discrepancies, including improper feature selection, insufficient data preprocessing, and suboptimal model hyperparameter settings. Based on the results of the discrepancy analysis, re-evaluate and select data features that are more representative or better reflect fault characteristics. Adjust the model's hyperparameters, including the number and depth of decision trees, and retrain the model using updated features and data. Through continuous iterations of training, prediction, evaluation, and tuning, the model's prediction accuracy and generalization ability can be gradually improved.

7. The method for automatically locating and isolating a distribution line fault according to claim 6, characterized in that: Step S4 specifically includes: S41: After receiving the fault signal from the intelligent fault indicator, the central processor first determines the location of the fault point and the fault type; S42: The central processor uses the topological information of the power distribution network to analyze and determine all isolation paths, taking into account the importance of power supply in the fault area, the stability of the network after the fault, and the number of affected users; S43: Based on the location of the fault point and all isolation paths, the central processor uses an optimization algorithm to calculate the best isolation solution, which aims to minimize the impact of the fault on the network and users, while ensuring rapid isolation of the fault area and maximum stability of power supply in non-fault areas.

8. The method for automatically locating and isolating a distribution line fault according to claim 7, characterized in that: After determining the best isolation scheme in step S4, the central processor issues instructions through wireless communication technology to control the corresponding intelligent circuit breakers or isolating switches in the distribution network to perform isolation operations, including remotely modifying the opening and closing states of the circuit breakers or switches to ensure that the fault area is quickly isolated. After the fault isolation operation is performed, the central processor will monitor the network load and power supply conditions outside the fault area, and adjust network operating parameters or start the backup power supply plan to maintain stable operation of the network.

9. The method for automatically locating and isolating a distribution line fault according to claim 8, characterized in that: Step S5 specifically includes: S51: The dynamic adaptive control module collects real-time data from each intelligent fault indicator in the power distribution network, including current, voltage, frequency and temperature information, as well as network topology change information after fault isolation; S52: The collected data will be used to update the state model of the distribution network in real time, which includes the current network load distribution, voltage level and possible power supply paths; S53: Based on the updated network state model, the dynamic adaptive control module uses the shortest path algorithm to calculate the optimal load distribution and current path; After determining the optimal load distribution and current path, the dynamic adaptive control module sends control instructions to the relevant intelligent circuit breakers, disconnectors and load management equipment through wireless communication technology, guiding the adjustment of operating status according to the optimization results, including adjusting the opening and closing of circuit breakers, changing the connection path of disconnectors or adjusting the load output of key nodes.

Images