Power distribution network fault source positioning method and device, electronic equipment and storage medium

Abstract

The invention provides a power distribution network fault source positioning method and device, electronic equipment and a storage medium. The method comprises the steps of collecting real-time operation data of a target power distribution network and performing data preprocessing; performing fault feature extraction on the real-time operation data after data preprocessing to obtain corresponding fault instability features; performing fault classification on the fault instability characteristics through a pre-trained deep neural network to obtain a corresponding fault classification result; and performing fault position judgment according to the fault classification result by adopting a test signal injection method to obtain a fault source positioning result of the target power distribution network. According to the power distribution network fault source positioning scheme provided by the invention, the power distribution network operation data is analyzed through the deep neural network, the fault of the power distribution network can be identified more accurately, the position of the fault source can be positioned more accurately in combination with the output of the S injection method and the deep neural network, the troubleshooting time is shortened, and the fault processing efficiency is improved. And the reliability of the power distribution network is improved.

Description

Technical Field

[0001] The present application relates to the technical field of power system monitoring, and in particular to a method, device, electronic device and storage medium for locating a fault source in a distribution network. Background Art

[0002] In modern power systems, the distribution network is a key link in power transmission, and its stability and reliability are crucial to ensuring the normal operation of social and economic activities. However, the complexity and variability of the distribution network have led to long-standing problems in the accuracy and speed of fault location. With the access of distributed power sources, the topology of the distribution network has become more complex, which further increases the difficulty of fault location.

[0003] Traditional distribution network fault location methods mainly include impedance method, traveling wave method and matrix algorithm. The impedance method relies on the impedance characteristics of the line to determine the fault location, but with the complexity of the distribution line, its location efficiency and accuracy gradually decrease. The traveling wave method realizes fault distance measurement by detecting the traveling wave signals at both ends of the line, but the dispersion problem during the propagation of the traveling wave requires complex signal processing technology to solve, which increases the complexity of the calculation. Although the matrix algorithm has a fast calculation speed, it has poor fault tolerance and is sensitive to signal distortion, which easily affects the efficiency of fault location. In addition, although optimization methods such as genetic algorithms have global search capabilities, redundant iterations may occur during the calculation process, affecting the accuracy and speed of fault location. In some areas with relatively backward economic development, manual fault location methods are still used. This method not only causes a waste of human resources, but also has low location efficiency, and the fault repair time is difficult to guarantee, and may even cause the fault range to expand. It can be seen that the existing distribution network fault source location technology has the problems of low location efficiency, low accuracy, poor adaptability and data instability, which reduces the reliability of the distribution network.

[0004] The preceding description is intended to provide general background information and does not necessarily constitute prior art. Summary of the invention

[0005] In response to the above technical problems, the present application provides a distribution network fault source locating method, device, electronic device and storage medium to solve the defects of low positioning efficiency and low accuracy in the prior art, accurately locate the position of the fault source, improve fault handling efficiency, and thus improve the reliability of the power system.

[0006] In order to solve the above technical problems, the present application provides a method for locating a fault source in a distribution network, comprising the following steps:

[0007] Collect real-time operation data of the target distribution network and perform data preprocessing;

[0008] Extract fault features from real-time operation data after data preprocessing to obtain corresponding fault instability features;

[0009] Fault classification is performed on the fault instability characteristics by using a pre-trained deep neural network to obtain a corresponding fault classification result;

[0010] A test signal injection method is used to determine the fault location according to the fault classification result to obtain a fault source location result of the target distribution network.

[0011] Furthermore, in some embodiments of the present application, the collecting of real-time operation data of the target distribution network and performing data preprocessing includes:

[0012] Installing sensors at monitoring points of the target distribution network;

[0013] The sensor monitors the operating state of the target power distribution network in real time to obtain real-time operating data, wherein the real-time operating data includes real-time voltage data, real-time current data, three-phase current imbalance data and power factor data;

[0014] The real-time operation data is preprocessed, and the data preprocessing includes data cleaning, denoising and standardization.

[0015] Further, in some embodiments of the present application, the extraction of fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features includes:

[0016] Extracting load rate characteristics based on the preprocessed real-time operation data using a load rate calculation formula to obtain load rate characteristics corresponding to the target distribution network;

[0017] Extracting three-phase imbalance characteristics according to the preprocessed real-time operation data using a three-phase imbalance calculation formula to obtain three-phase imbalance characteristics corresponding to the target distribution network;

[0018] A characteristic matrix is ​​constructed for the load rate characteristic and the three-phase imbalance characteristic to obtain a corresponding load rate characteristic matrix and a three-phase imbalance characteristic matrix.

[0019] Further, in some embodiments of the present application, the fault classification of the fault instability feature by a pre-trained deep neural network to obtain a corresponding fault classification result includes:

[0020] Converting the fault instability feature into a corresponding fault instability feature vector;

[0021] Inputting the fault instability feature vector into a pre-trained deep neural network and outputting a fault detection result;

[0022] After determining that a fault occurs in the target distribution network based on the fault detection result, the deep neural network is used to identify the fault type based on the fault detection result to obtain a corresponding fault classification result.

[0023] Further, in some embodiments of the present application, the method of using the test signal injection method to judge the fault location according to the fault classification result to obtain the fault source location result of the target distribution network includes:

[0024] Injecting at least one test signal at a preset position of the target distribution network using a test signal injection method;

[0025] Monitoring the test signal and generating a signal monitoring result;

[0026] Analyze the propagation path in the signal monitoring result to determine the area where the signal changes abnormally;

[0027] The fault location is determined based on the abnormal signal change area and the fault classification result to determine the fault source location result of the target distribution network.

[0028] Furthermore, in some embodiments of the present application, the training process of the deep neural network includes:

[0029] Obtain sample operation data of the distribution network and its corresponding sample labels; the sample labels include positive sample labels for characterizing the normal operation of the distribution network and negative sample labels for characterizing the failure of the distribution network, and the negative sample labels include labels of different fault types and locations;

[0030] Extracting sample fault instability features corresponding to the sample operation data through an initial deep neural network;

[0031] Preprocessing the sample fault instability characteristics;

[0032] Inputting the sample fault instability features into the initial deep neural network to obtain a sample fault classification result;

[0033] Adjusting the model parameters of the initial deep neural network based on the difference loss between the sample fault classification result and the sample label;

[0034] The initial deep neural network is continuously trained based on the adjusted model parameters until the training conditions are met, thereby obtaining a trained deep neural network.

[0035] Further, in some embodiments of the present application, after the fault location is determined according to the fault classification result by using the test signal injection method to obtain the fault source location result of the target distribution network, the method further includes:

[0036] Obtaining an on-site verification result of the fault source location result;

[0037] Based on the field verification results, the model parameters of the deep learning network and the algorithm parameters corresponding to the test signal injection method are adjusted.

[0038] Accordingly, the present application also provides a distribution network fault source location device, comprising:

[0039] Data acquisition module, used to collect real-time operation data of the target distribution network and perform data preprocessing;

[0040] The feature extraction module is used to extract fault features from the real-time operation data after data preprocessing to obtain the corresponding fault instability features;

[0041] A fault classification module is used to classify the fault instability characteristics through a pre-trained deep neural network to obtain a corresponding fault classification result;

[0042] The fault location module is used to determine the fault location according to the fault classification result by using a test signal injection method to obtain the fault source location result of the target distribution network.

[0043] The present application also provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of the distribution network fault source locating method as described above are implemented.

[0044] The present application also provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the distribution network fault source locating method as described above are implemented.

[0045] Implementing the embodiments of the present application has the following beneficial effects:

[0046] As described above, the present application provides a distribution network fault source location method, device, electronic device and storage medium, and the distribution network fault source location method includes: collecting real-time operation data of the target distribution network and performing data preprocessing; extracting fault features of the real-time operation data after data preprocessing to obtain corresponding fault instability features; classifying the fault instability features through a pre-trained deep neural network to obtain corresponding fault classification results; using a test signal injection method to judge the fault location according to the fault classification results to obtain the fault source location result of the target distribution network. The distribution network fault source location scheme proposed in the present application can accurately identify faults in the distribution network, including the existence of faults and the specific type of faults, by performing fault analysis on the distribution network operation data through a deep neural network, and can more accurately locate the location of the fault source by combining the test signal injection method and the output of the deep neural network, thereby shortening the troubleshooting time, reducing the operation and maintenance cost, and improving the fault handling efficiency, so that the distribution network can respond quickly when a fault occurs, reduce the power outage time, and improve the reliability of the power system. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] The drawings herein are incorporated into the specification and constitute a part of the specification, illustrate embodiments consistent with the present application, and together with the specification are used to explain the principles of the present application. In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments are briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor.

[0048] Figure 1 It is a flow chart of a method for locating a fault source in a distribution network provided in an embodiment of the present application;

[0049] Figure 2 is a schematic diagram of a test signal injection method provided in an embodiment of the present application;

[0050] Figure 3 It is a structural schematic diagram of a distribution network fault source locating device provided in an embodiment of the present application;

[0051] Figure 4 It is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application.

[0052] The realization of the purpose, functional features and advantages of this application will be further described in conjunction with the embodiments and with reference to the accompanying drawings. The above-mentioned drawings have shown clear embodiments of this application, which will be described in more detail later. These drawings and textual descriptions are not intended to limit the scope of the concept of this application in any way, but to illustrate the concept of this application to those skilled in the art by referring to specific embodiments. DETAILED DESCRIPTION

[0053] Exemplary embodiments will be described in detail herein, examples of which are shown in the accompanying drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application. Instead, they are merely examples of devices and methods consistent with some aspects of the present application as detailed in the appended claims.

[0054] It should be noted that, in this article, the terms "include", "comprises" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also includes other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the sentence "includes a ..." does not exclude the existence of other identical elements in the process, method, article or device including the element. In addition, components, features, and elements with the same name in different embodiments of the present application may have the same meaning or different meanings, and their specific meanings need to be determined by their explanation in the specific embodiment or further combined with the context of the specific embodiment.

[0055] It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0056] In the subsequent description, the suffixes such as "module", "component" or "unit" used to represent elements are only used to facilitate the description of the present application, and have no specific meanings. Therefore, "module", "component" or "unit" can be used in a mixed manner.

[0057] The complexity and variability of the distribution network have led to long-standing problems in the accuracy and speed of fault location. With the access of distributed power sources, the topology of the distribution network has become more complex, and the effectiveness of traditional fault location methods in practical applications has decreased. New technologies are needed to improve the accuracy and efficiency of fault location. Although the existing distribution network fault source location technology has made certain progress, it still has some shortcomings and limitations: Traditional impedance method: This method is simple and low-cost, but as the distribution lines become more and more complex and the number of branches increases, its location efficiency and accuracy are low. Traveling wave method: Although the traveling wave signals detected at both ends of the line can be used to measure the distance of short-circuit faults, the dispersion problem existing in the propagation of traveling waves needs to be solved by wavelet transform and other methods, which increases the complexity of the calculation. Matrix algorithm: The principle is simple and the calculation speed is fast, but the fault tolerance is poor. It is sensitive to the situation with distorted signals and is easy to affect the final fault location efficiency. Genetic algorithm: It has global search capabilities, but redundant iterations may occur near the end of the calculation, which affects the accuracy and speed of fault location and reduces the solution efficiency. Manual fault location: In some areas with relatively backward economic development, manual fault location is still used, which not only wastes human resources, but also has low location efficiency, makes it difficult to guarantee the time for fault repair, and may even lead to the expansion of the fault range. Therefore, developing a new fault source location method to improve the accuracy and efficiency of fault location, reduce operation and maintenance costs, and improve the reliability and power supply quality of the power grid is a technical problem that needs to be solved in the field of power systems.

[0058] In response to the above technical problems, the embodiments of the present application provide a distribution network fault source location method, device, electronic device and storage medium, which analyze the distribution network operation data through a deep neural network, can more accurately identify the faults of the distribution network, and combine the S injection method and the output of the deep neural network to more accurately locate the fault source, shorten the fault detection time, improve the fault handling efficiency, and thus improve the reliability of the distribution network.

[0059] The embodiments of the present application provide a method, device, electronic device and storage medium for locating a distribution network fault source. The distribution network fault source locating device can be deployed in a terminal or a server, the server can include an independently operated server or a distributed server, or a server cluster composed of multiple servers, and the terminal can include a mobile phone, a tablet computer or a personal computer (PC).

[0060] It should be noted that the order of description of the following embodiments is not intended to limit the priority order of the embodiments.

[0061] The present application provides a method for locating a distribution network fault source, comprising: collecting real-time operating data of a target distribution network and performing data preprocessing; extracting fault features from the real-time operating data after data preprocessing to obtain corresponding fault instability features; classifying the fault instability features through a pre-trained deep neural network to obtain corresponding fault classification results; and using a test signal injection method to determine the fault location based on the fault classification results to obtain a fault source locating result of the target distribution network.

[0062] See also Figure 1 , Figure 1 1 is a flow chart of a method for locating a distribution network fault source provided in an embodiment of the present application. The method for locating a distribution network fault source provided in this embodiment can be executed in a distribution network, and the method for locating a distribution network fault source can specifically include the following steps:

[0063] S1. Collect real-time operation data of the target distribution network and perform data preprocessing;

[0064] Specifically, for step S1, sensors are installed at key nodes of the distribution network in advance to monitor and collect real-time operation data of the target distribution network. The real-time operation data includes key electrical parameters such as voltage, current, power factor, and three-phase imbalance, which serve as the basis for fault detection and location. The collected raw data is cleaned to remove outliers and erroneous readings, and then the noise is removed by filtering technology. Finally, the data is normalized to eliminate the dimensional differences of data collected by different sensors and at different times, so as to prepare data for subsequent feature extraction and model training.

[0065] S2. Extract fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features;

[0066] Specifically, for step S2, fault instability features related to the fault, such as load rate and three-phase imbalance, are extracted from the preprocessed data. The fault instability features can reflect the operating state of the distribution network and are associated with the fault type.

[0067] S3. Fault classification is performed on the fault instability characteristics through a pre-trained deep neural network to obtain the corresponding fault classification results;

[0068] Specifically, for step S3, the extracted fault features and known fault type labels are used to train a deep neural network. The deep neural network builds a fault recognition model by learning the data features under normal and fault states, and the feature vectors of the real-time monitoring data are input into the trained deep neural network, and the model outputs the fault detection results and fault types.

[0069] S4. Use the test signal injection method to determine the fault location according to the fault classification result, and obtain the fault source location result of the target distribution network;

[0070] Specifically, for step S4, a test signal injection method is used to inject a test signal at a preset position of the distribution network according to the fault classification result, monitor the propagation of the signal, analyze the area of ​​abnormal signal change, and combine the output of the deep learning model to determine the precise location of the fault source.

[0071] It can be seen that the distribution network fault source location method provided in this embodiment improves the accuracy of fault detection and reduces the possibility of false alarms and missed alarms through the learning and identification of fault characteristics by the deep learning model; combined with the test signal injection method, it can accurately locate the fault source, providing important information for rapid response and maintenance, thereby reducing power outage time and maintenance costs; by timely discovering and handling faults, it improves the operational reliability of the distribution network and ensures the continuity and stability of power supply. In addition, the entire fault detection and location process is automated and intelligent, reducing dependence on manual intervention and improving operation and maintenance efficiency.

[0072] Furthermore, in some embodiments, step S1 “collecting real-time operation data of the target distribution network and performing data preprocessing” may specifically include:

[0073] S1. Install sensors at monitoring points of the target distribution network;

[0074] S2. Monitor the operating status of the target distribution network in real time through sensors to obtain real-time operating data, which includes real-time voltage data, real-time current data, three-phase current imbalance data and power factor data;

[0075] S3. Perform data preprocessing on the real-time operation data, which includes data cleaning, denoising and standardization.

[0076] Specifically, for step S1, the main purpose is to collect the real-time operation data of the distribution network and perform data preprocessing. The specific process is as follows: First, install various types of sensors at key locations of the distribution network, such as voltage transformers (VT), current transformers (CT) and smart meters, to ensure that the operation status of the distribution network can be fully monitored. The installed sensors should have high precision and high reliability to ensure the accuracy of the data; then, collect the electrical parameters of the distribution network in real time through the installed sensors, including but not limited to voltage, current, three-phase current imbalance and power factor. These data are crucial for analyzing the operation status of the distribution network and identifying potential faults. Finally, perform data preprocessing on the collected real-time operation data, including removing outliers and erroneous readings in the data to avoid the impact of sensor failure or external interference; apply filtering technology (such as low-pass filters) to reduce noise in the data and extract useful signal features; standardize the data, such as normalization or Z-score standardization, so that the data has a uniform scale and distribution, which is convenient for subsequent feature extraction and model training.

[0077] This embodiment improves the quality and reliability of data through data preprocessing, providing a solid foundation for subsequent fault feature extraction and analysis; the cleaned and denoised data more realistically reflects the operating status of the distribution network, helping to improve the accuracy of fault detection; the standardized data enables the deep learning model to learn and identify fault characteristics more effectively, improving the training efficiency and prediction effect of the model; through precise data collection and preprocessing, the system can operate more stably, improving the reliability of distribution network fault detection and positioning; accurate data collection and preprocessing reduce false alarms and missed alarms, thereby reducing the costs caused by unnecessary inspections and repairs.

[0078] Further, in some embodiments, step S2 of “extracting fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features” may specifically include:

[0079] S21. Extract the load rate characteristics according to the preprocessed real-time operation data through the load rate calculation formula to obtain the load rate characteristics corresponding to the target distribution network;

[0080] S22. Extract three-phase imbalance characteristics according to the preprocessed real-time operation data through the three-phase imbalance calculation formula to obtain the three-phase imbalance characteristics corresponding to the target distribution network;

[0081] S23. Construct a characteristic matrix for the load rate characteristics and the three-phase imbalance characteristics to obtain a corresponding load rate characteristic matrix and a three-phase imbalance characteristic matrix.

[0082] Specifically, for step S2, the fault instability characteristics in this embodiment include load rate characteristics and three-phase imbalance characteristics, wherein the load rate characteristic is an important operating parameter in the distribution network, reflecting the utilization efficiency of the power equipment in the distribution network. The calculation of the load rate is usually based on the ratio of active power and reactive power to the rated capacity of the distribution transformer. Through real-time monitoring data, the load rate of each monitoring point can be calculated, and these data will be used for subsequent fault analysis. For three-phase imbalance, it is used as an indicator to measure whether the three-phase current of the distribution network is balanced. Ideally, the three-phase current should be balanced. If imbalance occurs, it may be caused by some faults or asymmetric loads in the system. The calculation of three-phase imbalance can be achieved by comparing the maximum and minimum values ​​of the three-phase current.

[0083] The extracted load rate and three-phase imbalance features are organized into feature matrices that conform to the input format of the deep learning model, and these matrices will be used as inputs to the deep learning model. The construction of feature matrices helps to structure complex data, making it more suitable for pattern recognition and machine learning algorithms.

[0084] In a specific embodiment, the load rate calculation formula can be expressed as:

[0085]

[0086] Where P 有 , P 无 They represent active power and reactive power respectively; C represents the rated capacity of the distribution transformer.

[0087] The three-phase unbalance calculation formula can be expressed as:

[0088]

[0089] In the formula, I 最大项 Indicates the maximum phase current; I 最小项 Indicates the minimum phase current. When the three-phase imbalance exceeds a preset threshold (eg, 25%), it indicates that three-phase imbalance occurs.

[0090] In a specific embodiment, when analyzing the distribution network data, the distribution network stable operation element is set as E 0 , there is only noise signal; suppose the distribution network fault is E 1 , there are noise signals and fault information. Assume that the acquisition time is t 0 , E at that time 0 and E 1 It can be expressed as:

[0091]

[0092] In the formula, s(t 0) is a Gaussian noise signal; h 1 (t 0 ) is the impulse response signal; x 0 (t 0 ) is the initial signal. 0 In this case, each element in the matrix is ​​a random variable and is independently and uniformly distributed.

[0093] This embodiment can more accurately identify abnormal conditions and potential faults in the distribution network by extracting key features such as load rate and three-phase imbalance. These features provide rich information for the deep learning model, which helps to improve the accuracy of fault location and thus find the source of the fault more quickly. By analyzing the load rate, the load distribution of the distribution network can be better understood, which helps to optimize the allocation and scheduling of power resources. The monitoring and analysis of three-phase imbalance helps to promptly discover and solve imbalance problems, thereby improving the stability and reliability of the distribution network.

[0094] Further, in some embodiments, step S3 "fault classification of the fault instability characteristics by using a pre-trained deep neural network to obtain a corresponding fault classification result" may specifically include:

[0095] S31. Converting the fault instability feature into a corresponding fault instability feature vector;

[0096] S32. Input the fault instability feature vector into a pre-trained deep neural network and output the fault detection result;

[0097] S33. After determining that a fault occurs in the target distribution network based on the fault detection results, a deep neural network is used to identify the fault type based on the fault detection results to obtain a corresponding fault classification result.

[0098] Specifically, for step S3, the extracted fault instability features (such as load rate, three-phase imbalance, etc.) are first converted into numerical feature vectors. This step ensures that the feature data can be processed by the deep neural network. The feature vector is a point in a high-dimensional space, and each dimension represents a feature. Then, the fault instability feature vector is input into a pre-trained deep neural network. This network can contain multiple hidden layers and can learn the complex mapping relationship between feature vectors and fault types. Finally, the input feature vector is processed by the deep neural network, and fault detection is first performed, that is, it is determined whether there is a fault in the distribution network. If a fault is detected, the network will further identify the fault type, such as distinguishing different fault types such as single-phase grounding and two-phase short circuit.

[0099] In addition, during the training phase, the deep neural network is trained using labeled fault data, and the network weights are adjusted through the back-propagation algorithm to minimize the difference between the predicted results and the actual fault type. The training process may include multiple iterations until the network performance meets the predetermined criteria.

[0100] In a specific embodiment, the obtained normalized data is input into a deep neural network in a matrix format to classify the corresponding faults. The forward data transmission function of deep learning is:

[0101]

[0102] In the formula, δ represents the activation function; n represents the number of neuron layers; w n represents the weight parameter matrix; x n-1 Represents input; represents the bias parameter matrix.

[0103] The deep neural network of this embodiment can learn the fault pattern from complex data, significantly improving the accuracy of fault detection. Once a fault is detected, the deep neural network can quickly classify the fault type and provide decision support for rapid response; the automated fault detection and classification process reduces the reliance on manual intervention and improves the automation level of the distribution network; accurate fault detection and classification helps to quickly locate the source of the fault, thereby shortening the repair time and reducing the impact of power outages on users; by timely discovering and handling faults, the operational reliability of the distribution network is improved and the stability of power supply is guaranteed; automated fault detection and classification reduces the need for manual inspections and reduces operation and maintenance costs. At the same time, fast and accurate fault handling reduces economic losses caused by faults.

[0104] Further, in some embodiments, step S4 "using the test signal injection method to determine the fault location according to the fault classification result to obtain the fault source location result of the target distribution network" may specifically include:

[0105] S41. Inject at least one test signal into a preset position of the target distribution network using a test signal injection method;

[0106] S42 monitors the test signal and generates a signal monitoring result;

[0107] S43. Analyze the propagation path in the signal monitoring results to determine the area where the signal changes abnormally;

[0108] S44. The fault location is judged based on the abnormal signal change area and the fault classification result, and the fault source location result of the target distribution network is determined.

[0109] Specifically, for step S4, a test signal of known characteristics is first injected into a preset key position in the distribution network. The injected test signal usually has specific frequency and phase characteristics so that it can be tracked and analyzed. Then, the injected test signal is monitored at multiple locations in the distribution network. By monitoring the propagation and change of the signal, data about the propagation of the signal in the power grid can be collected. Next, the propagation path of the test signal in the distribution network is analyzed, especially focusing on parameters such as signal attenuation, phase change and propagation time. Changes in these parameters can reflect abnormal conditions in the power grid. Then, combined with the fault classification results provided by the deep learning model and the propagation path analysis of the test signal, the approximate area where the fault occurs is determined. The area where the signal changes abnormally usually indicates an area closely related to the fault location. Finally, in the determined abnormal area, the specific changes of the signal, such as local attenuation increase or phase mutation, are further analyzed to accurately locate the fault point.

[0110] Specifically, the S injection process is as follows: Figure 2 As shown, with the support of supervised training algorithm, the source location is found by combining S injection method. Figure 2 When a power supply fault occurs in the feeder, the signal injector injects a fault location signal with a different power supply frequency into the secondary end of the voltage transformer. The fault location signal is connected to one side of the transformer to form a loop. Figure 2 The middle dotted line represents the fault location signal path. The fault location is determined based on the fault location trajectory.

[0111] This embodiment can more accurately locate the source of the fault and reduce the troubleshooting time by combining the fault classification results of the deep learning model and the propagation path analysis of the test signal; accurate fault location enables the maintenance team to quickly reach the fault site, reducing the fault response time and improving maintenance efficiency; fast and accurate fault location helps to shorten the power outage time, reduce the impact on users, and improve service quality; by timely discovering and repairing faults, the operating reliability of the distribution network is improved and the risk of power outages caused by faults is reduced; the combination of the test signal injection method and the deep learning model provides data-driven decision support, which helps to optimize the operation and maintenance strategy of the power grid. In addition, this method can adapt to different operating states and fault types of the distribution network, so it has good versatility and adaptability.

[0112] Furthermore, in some embodiments, the training process of the deep neural network may specifically include:

[0113] Obtain sample operation data of the distribution network and its corresponding sample labels; the sample labels include positive sample labels for characterizing the normal operation of the distribution network and negative sample labels for characterizing the failure of the distribution network, and the negative sample labels include labels of different fault types and locations;

[0114] Extract the sample fault instability features corresponding to the sample operation data through the initial deep neural network;

[0115] Preprocessing the sample fault instability characteristics;

[0116] Input the sample fault instability features into the initial deep neural network to obtain the sample fault classification results;

[0117] Adjust the model parameters of the initial deep neural network based on the difference loss between the sample fault classification results and the sample labels;

[0118] The initial deep neural network is trained continuously based on the adjusted model parameters until the training conditions are met to obtain a trained deep neural network.

[0119] Specifically, for the deep neural network in this embodiment, this embodiment also provides a training method for a deep neural network, and the specific process is as follows: First, collect data of the distribution network under different operating conditions, including data under normal operation and various fault conditions. Assign corresponding labels to each set of data, normal operation as positive samples, and different fault types and locations as negative samples. Next, use the initial deep neural network to extract fault instability features from the sample data. These features can reflect the operating state and potential fault modes of the distribution network. The extracted features are further preprocessed, such as normalization or standardization, to ensure that the feature data is suitable for the training of the deep learning model. Then, the preprocessed features are input into the deep neural network for preliminary fault classification. According to the difference (loss) between the classification result and the actual label, the network parameters are adjusted using the back propagation algorithm. Finally, through the iterative training process, the network parameters are continuously adjusted to optimize the model performance until the model can accurately classify the fault type and location. During the training process, the performance of the model will also be evaluated on independent validation sets and test sets to ensure that the model has good generalization ability and can process new data.

[0120] In a specific embodiment, this embodiment uses a supervised training algorithm to train a deep neural network. In a multi-classification problem, the corresponding logarithmic function is used as the loss number, that is:

[0121]

[0122] In the formula, m is the number of samples; y is j (x) is the output of the deep neural network, y i (x) is the prediction output of the deep neural network.

[0123] The calculation formula for the gradient of each parameter of the back propagation algorithm in the deep neural network is:

[0124]

[0125] In the formula, k represents the number of updates; S loss represents the partial derivative function; Represents the update coefficient. The square gradient moving average of the parameter at the kth update is calculated as:

[0126] d (k) =μd (k-1) +(1-μ)dT (k)

[0127] Where, d (k-1) represents the moving average of the square gradient of the parameter at the k-1th update, and μ represents the gradient decay rate. By using the loss function to back propagate, the changes of each parameter are obtained and corrected, providing accurate parameters for fault location.

[0128] This embodiment significantly improves the accuracy of fault detection and classification through the training and optimization of the deep learning model; after sufficient training and verification, the model has good generalization ability and can adapt to different distribution network environments and fault types; the trained deep neural network can automatically learn and identify fault modes from large amounts of data, reducing the need for manual analysis; accurate fault detection and classification help to quickly locate the source of the fault, shorten the repair time, and improve the operation and maintenance efficiency of the distribution network.

[0129] Furthermore, in some embodiments, after step S4 "using the test signal injection method to determine the fault location according to the fault classification result to obtain the fault source location result of the target distribution network", it may also include:

[0130] S51. Obtaining on-site verification results of fault source location results;

[0131] S52. Adjust the model parameters of the deep learning network and the algorithm parameters corresponding to the test signal injection method based on the field verification results.

[0132] Specifically, this embodiment also provides a parameter adjustment step. After obtaining the preliminary positioning result of the fault source, a technician is dispatched to the located fault point for on-site verification, including checking the actual state of the distribution network and confirming whether the fault type and location are consistent with the model prediction. The results of the on-site verification are compared and analyzed with the results of the model prediction. If a difference is found, the cause needs to be further analyzed, which may be inaccurate data collection, insufficient model training, or improper parameter setting of the test signal injection method. According to the results of the on-site verification, the model parameters of the deep learning network are adjusted. This may include adjusting the network structure, optimizing the algorithm, or retraining the model to improve the accuracy and reliability of the model. At the same time, the algorithm parameters of the test signal injection method are adjusted, such as changing the frequency, amplitude, or injection position of the test signal to optimize the propagation characteristics and monitoring effect of the signal. Finally, the adjusted model and algorithm are applied to the actual distribution network, and new data continues to be collected to conduct a new round of fault detection and location to verify the adjustment effect, and further optimize as needed.

[0133] Through field verification and adjustment of model parameters, this embodiment can significantly improve the accuracy of fault location and reduce misjudgments and missed judgments; continuous optimization and adjustment enable the system to adapt to changes in the distribution network, including new fault modes and changes in the grid structure; field verification and model adjustment form a closed-loop control system, ensuring continuous improvement and optimization of fault detection and location methods.

[0134] In a specific embodiment, the method for locating the source of a distribution network fault provided in this embodiment constructs a binary model of a distribution network fault and analyzes the instability characteristics of the distribution network fault data. The obtained normalized data is input into a deep neural network in a matrix format to obtain classification indicators of load rate and three-phase imbalance. The relationship between the event information pool and the fault analysis is analyzed to obtain the fault classification result. A supervised training algorithm of deep learning is used in combination with the S injection method to find the source fault location.

[0135] Different from the traditional distribution network fault source location technology, the method disclosed in this embodiment is to obtain the load rate and classify the three-phase imbalance index based on the instability analysis of the distribution network fault data. Analyze the relationship between the event information pool and the fault judgment to obtain the fault classification result. The distribution network fault source is located using a supervised training algorithm of deep learning. Before the general basic module such as the distribution network fault data analysis module, a distribution network fault data acquisition module is pre-placed. The acquisition module collects the elements of the distribution network fault information of any user in any distribution network environment. The collected elements include but are not limited to the user voltage, current, three-phase imbalance characteristics, etc.

[0136] The present embodiment provides a distribution network fault source location method based on artificial intelligence to solve the problem of determining the fault section where the fault point is located, so as to isolate the fault and restore power supply to non-fault areas. This can not only improve the efficiency and accuracy of fault handling, but also reduce operation and maintenance costs, and improve the reliability and power supply quality of the entire power grid.

[0137] In summary, the present embodiment provides a method for locating the fault source of a distribution network, including: collecting real-time operation data of the target distribution network and performing data preprocessing; extracting fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features; classifying the fault instability features through a pre-trained deep neural network to obtain corresponding fault classification results; using a test signal injection method to judge the fault location according to the fault classification results to obtain the fault source location result of the target distribution network. It can be seen that the distribution network fault source location method provided by the present embodiment, through the deep neural network, performs fault analysis on the distribution network operation data, can accurately identify the faults in the distribution network, including the existence of the fault and the specific type of the fault, and combined with the test signal injection method and the output of the deep neural network, can more accurately locate the location of the fault source, thereby shortening the troubleshooting time, reducing the operation and maintenance cost, and improving the fault handling efficiency, so that the distribution network can respond quickly when a fault occurs, reduce the power outage time, and improve the reliability of the power system.

[0138] In order to better implement the distribution network fault source location method of the embodiment of the present application, the embodiment of the present application also provides a distribution network fault source location device. The meanings of the terms are the same as those in the above distribution network fault source location method, and the specific implementation details can refer to the description in the method embodiment.

[0139] See also Figure 3 , Figure 3 A schematic diagram of the structure of a distribution network fault source location device provided in an embodiment of the present application, wherein the distribution network fault source location device may specifically include a data acquisition module 201, a feature extraction module 202, a fault classification module 203 and a fault location module 204, which may be specifically as follows:

[0140] The data acquisition module 201 is used to collect the real-time operation data of the target distribution network and perform data preprocessing;

[0141] The feature extraction module 202 is used to extract fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features;

[0142] A fault classification module 203 is used to classify the fault instability characteristics through a pre-trained deep neural network to obtain a corresponding fault classification result;

[0143] The fault location module 204 is used to determine the fault location according to the fault classification result by using the test signal injection method to obtain the fault source location result of the target distribution network.

[0144] In a specific embodiment, the distribution network fault source locating device provided in this embodiment can also be divided into the following modules:

[0145] The distribution network fault data analysis module is the core module for locating the distribution network fault source. The entire architecture needs to rely on this module to complete the data analysis. The distribution network fault data analysis module mainly has two components, which need to complete the collection of distribution network fault data elements and use algorithms to analyze the collected elements.

[0146] The distribution network fault data classification module inputs the obtained structured data into the deep neural network in matrix format to classify the corresponding faults. The faults are mainly caused by load rate and three-phase imbalance, and it is necessary to calculate the distribution network operation status under different indicators. Combining the above two indicators, the type is analyzed from the perspective of the event information pool. When the interruption area in the event database covers the distribution network, the fault analysis and search program will be terminated; through the correlation analysis of the distribution network and the branch, the status of the distribution network and the branch are extracted respectively to determine the distribution network fault.

[0147] The distribution network fault source location module based on deep learning conducts supervised training on the distribution network fault classification data, and locates the distribution network fault through supervised training algorithms and fault location judgment.

[0148] To sum up, the distribution network fault source locating device provided in this embodiment performs fault analysis on the distribution network operation data through a deep neural network, and can accurately identify faults in the distribution network, including the presence or absence of a fault and the specific type of the fault, and combines the test signal injection method and the output of the deep neural network to more accurately locate the fault source, thereby shortening the troubleshooting time, reducing operation and maintenance costs, and improving fault handling efficiency, so that the distribution network can respond quickly when a fault occurs, reduce power outage time, and improve the reliability of the power system.

[0149] In addition, the present application also provides an electronic device, such as Figure 4 , which shows a schematic diagram of the structure of an electronic device involved in an embodiment of the present application. Specifically, the electronic device may include a processor 301 with one or more processing cores, a memory 302 with one or more computer-readable storage media, a power supply 303, an input unit 304 and other components. Those skilled in the art can understand that Figure 4 The electronic device structure shown in the figure does not constitute a limitation on the electronic device, and may include more or fewer components than shown in the figure, or combine certain components, or arrange the components differently.

[0150] The processor 301 is the control center of the electronic device. It uses various interfaces and lines to connect various parts of the entire electronic device. By running or executing software programs and / or modules stored in the memory 302, and calling data stored in the memory 302, it executes various functions of the electronic device and processes data, thereby monitoring the electronic device as a whole. Optionally, the processor 301 may include one or more processing cores; preferably, the processor 301 may integrate an application processor and a modem processor, wherein the application processor mainly processes the operating system, user interface, and application programs, and the modem processor mainly processes wireless communications. It is understandable that the above-mentioned modem processor may not be integrated into the processor 301.

[0151] The memory 302 can be used to store software programs and modules. The processor 301 executes various functional applications and a method for locating a fault source in a power distribution network by running the software programs and modules stored in the memory 302. The memory 302 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system, an application required for at least one function (such as a sound playback function, an image playback function, etc.), etc.; the data storage area may store data created according to the use of the electronic device, etc. In addition, the memory 302 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other volatile solid-state storage devices. Accordingly, the memory 302 may also include a memory controller to provide the processor 301 with access to the memory 302.

[0152] The electronic device also includes a power supply 303 for supplying power to each component. Preferably, the power supply 303 can be logically connected to the processor 301 through a power management system, so that the power management system can manage charging, discharging, power consumption and other functions. The power supply 303 can also include one or more DC or AC power supplies, recharging systems, power failure detection circuits, power converters or inverters, power status indicators and other arbitrary components.

[0153] The electronic device may further include an input unit 304, which may be used to receive input digital or character information and generate keyboard, mouse, joystick, optical or trackball signal input related to user settings and function control.

[0154] Although not shown, the electronic device may further include a display unit, etc., which will not be described in detail herein. Specifically in this embodiment, the processor 301 in the electronic device will load the executable files corresponding to the processes of one or more application programs into the memory 302 according to the following instructions, and the processor 301 will run the application programs stored in the memory 302, thereby realizing various functions, as follows:

[0155] The real-time operation data of the target distribution network is collected and preprocessed; fault features are extracted from the real-time operation data after data preprocessing to obtain the corresponding fault instability features; the fault instability features are classified by a pre-trained deep neural network to obtain the corresponding fault classification results; the test signal injection method is used to determine the fault location according to the fault classification results to obtain the fault source location result of the target distribution network.

[0156] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0157] The embodiment of the present application performs fault analysis on the distribution network operation data through a deep neural network, and can accurately identify faults in the distribution network, including the presence or absence of a fault and the specific type of the fault. In combination with the test signal injection method and the output of the deep neural network, the location of the fault source can be more accurately located, thereby shortening the troubleshooting time, reducing operation and maintenance costs, and improving fault handling efficiency, so that the distribution network can respond quickly when a fault occurs, reducing power outage time, and improving the reliability of the power system.

[0158] A person of ordinary skill in the art will appreciate that all or part of the steps in the various methods of the above embodiments may be completed by instructions, or by controlling related hardware through instructions. The instructions may be stored in a computer-readable storage medium and loaded and executed by a processor.

[0159] To this end, an embodiment of the present application provides a storage medium in which a plurality of instructions are stored, and the instructions can be loaded by a processor to execute the steps in any one of the distribution network fault source location methods provided in the embodiments of the present application. For example, the instructions can execute the following steps:

[0160] The real-time operation data of the target distribution network is collected and preprocessed; fault features are extracted from the real-time operation data after data preprocessing to obtain the corresponding fault instability features; the fault instability features are classified by a pre-trained deep neural network to obtain the corresponding fault classification results; the test signal injection method is used to determine the fault location according to the fault classification results to obtain the fault source location result of the target distribution network.

[0161] The specific implementation of the above operations can be found in the previous embodiments, which will not be described in detail here.

[0162] The storage medium may include: a read-only memory (ROM), a random access memory (RAM), a disk or an optical disk, etc. Since the instructions stored in the storage medium can execute the steps in any of the distribution network fault source location methods provided in the embodiments of the present application, the beneficial effects that can be achieved by any of the distribution network fault source location methods provided in the embodiments of the present application can be achieved. For details, please refer to the previous embodiments and will not be repeated here.

[0163] The above is a detailed introduction to a distribution network fault source locating method, device, electronic device and storage medium provided in an embodiment of the present application. Specific examples are used in this article to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method of the present application and its core idea; at the same time, for technical personnel in this field, according to the idea of ​​the present application, there will be changes in the specific implementation method and application scope. In summary, the content of this specification should not be understood as a limitation on the present application.

Claims

1. A method for locating a fault source in a distribution network, characterized in that: The steps include: Collect real-time operation data of the target distribution network and perform data preprocessing; Extract fault features from real-time operation data after data preprocessing to obtain corresponding fault instability features; Fault classification is performed on the fault instability characteristics by using a pre-trained deep neural network to obtain a corresponding fault classification result; A test signal injection method is used to determine the fault location according to the fault classification result to obtain a fault source location result of the target distribution network.

2. The method for locating a fault source in a distribution network according to claim 1, characterized in that: The collecting of real-time operation data of the target distribution network and performing data preprocessing includes: Installing sensors at monitoring points of the target distribution network; The sensor monitors the operating state of the target power distribution network in real time to obtain real-time operating data, wherein the real-time operating data includes real-time voltage data, real-time current data, three-phase current imbalance data and power factor data; The real-time operation data is preprocessed, and the data preprocessing includes data cleaning, denoising and standardization.

3. The method for locating a fault source in a distribution network according to claim 1, characterized in that: The extracting of fault features from the real-time operation data after data preprocessing to obtain corresponding fault instability features includes: Extracting load rate characteristics based on the preprocessed real-time operation data using a load rate calculation formula to obtain load rate characteristics corresponding to the target distribution network; Extracting three-phase imbalance characteristics according to the preprocessed real-time operation data using a three-phase imbalance calculation formula to obtain three-phase imbalance characteristics corresponding to the target distribution network; A characteristic matrix is ​​constructed for the load rate characteristic and the three-phase imbalance characteristic to obtain a corresponding load rate characteristic matrix and a three-phase imbalance characteristic matrix.

4. The method for locating a fault source in a distribution network according to claim 1, characterized in that: The method of performing fault classification on the fault instability characteristics by using a pre-trained deep neural network to obtain a corresponding fault classification result includes: Converting the fault instability feature into a corresponding fault instability feature vector; Inputting the fault instability feature vector into a pre-trained deep neural network and outputting a fault detection result; After determining that a fault occurs in the target distribution network based on the fault detection result, the deep neural network is used to identify the fault type based on the fault detection result to obtain a corresponding fault classification result.

5. The method for locating a fault source in a distribution network according to claim 1, characterized in that: The method of using the test signal injection method to determine the fault location according to the fault classification result to obtain the fault source location result of the target distribution network includes: Injecting at least one test signal at a preset position of the target distribution network using a test signal injection method; Monitoring the test signal and generating a signal monitoring result; Analyze the propagation path in the signal monitoring result to determine the area where the signal changes abnormally; The fault location is determined based on the abnormal signal change area and the fault classification result to determine the fault source location result of the target distribution network.

6. The method for locating a fault source in a distribution network according to claim 1, characterized in that: The training process of the deep neural network includes: Obtain sample operation data of the distribution network and its corresponding sample labels; the sample labels include positive sample labels for characterizing the normal operation of the distribution network and negative sample labels for characterizing the failure of the distribution network, and the negative sample labels include labels of different fault types and locations; Extracting sample fault instability features corresponding to the sample operation data through an initial deep neural network; Preprocessing the sample fault instability characteristics; Inputting the sample fault instability features into the initial deep neural network to obtain a sample fault classification result; Adjusting the model parameters of the initial deep neural network based on the difference loss between the sample fault classification result and the sample label; The initial deep neural network is continuously trained based on the adjusted model parameters until the training conditions are met, thereby obtaining a trained deep neural network.

7. The method for locating a fault source in a distribution network according to claim 1, characterized in that: After the fault location is determined according to the fault classification result by using the test signal injection method to obtain the fault source location result of the target distribution network, the method further includes: Obtaining an on-site verification result of the fault source location result; Based on the field verification results, the model parameters of the deep learning network and the algorithm parameters corresponding to the test signal injection method are adjusted.

8. A distribution network fault source location device, characterized in that: include: Data acquisition module, used to collect real-time operation data of the target distribution network and perform data preprocessing; The feature extraction module is used to extract fault features from the real-time operation data after data preprocessing to obtain the corresponding fault instability features; A fault classification module is used to classify the fault instability characteristics through a pre-trained deep neural network to obtain a corresponding fault classification result; The fault location module is used to determine the fault location according to the fault classification result by using a test signal injection method to obtain the fault source location result of the target distribution network.

9. An electronic device comprising a memory and a processor, wherein a computer program is stored in the memory, wherein: When the processor executes the computer program, the steps of the method for locating a fault source in a distribution network as claimed in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method for locating a fault source in a power distribution network as claimed in any one of claims 1 to 7 are implemented.

Images