Power distribution network fault section positioning method and system

Abstract

The invention discloses a power distribution network fault section positioning method and system. The method comprises the steps of establishing a target power distribution network model, and performing first preset section division based on the target power distribution network model; obtaining first target data of different short circuit faults at different section positions after the first preset section is divided; performing first preprocessing on the first target data to obtain a second target data set; and training a first power distribution network fault section positioning model based on the second target data set, and performing power distribution network fault section positioning according to the first power distribution network fault section positioning model. According to the power distribution network fault section positioning method and system, powerful technical support is provided for stable operation of a power system, power failure time is shortened, power supply reliability is improved, and the method and system are of great significance in guaranteeing stable operation of society and economy.

Description

Technical Field

[0001] The present invention relates to the technical field of fault section positioning in a distribution network, and in particular to a method and system for positioning a fault section in a distribution network. Background Art

[0002] As the last link in the power system to transmit electric energy to users, the reliable operation of the distribution network is crucial to the stable development of the entire society. However, in the actual operation of the distribution network, various faults often occur due to various factors such as natural disasters such as lightning strikes and strong winds, insulation and equipment aging due to moisture, and manual operation errors. Such as single-phase short circuit to ground fault, two-phase short circuit fault, etc. These faults will not only cause power outages and affect the normal power consumption of users, but may also cause damage to power grid equipment, causing huge economic losses and safety hazards. Therefore, it is of great significance to quickly and accurately identify the faulty section of the distribution network, take measures to restore power supply, and ensure the stable operation of the power system.

[0003] Some scholars have studied the fault location method based on traveling waves. By measuring the time difference of traveling waves arriving at different locations and combining the propagation speed of power lines, the location of the fault point can be calculated. However, the cost of traveling wave measurement equipment is high and its application scenarios are relatively limited. Some scholars have studied the fault location method based on impedance measurement, which locates the fault by measuring the impedance value from the power source to the impedance point. This method is simple and feasible, but its accuracy is greatly affected by line parameters. However, when a large number of distributed generators are connected to the distribution network, the direction, size and characteristics of the current will change in the event of a fault, resulting in the traditional distribution network fault location method no longer being applicable. Therefore, the traditional distribution network section location method has been unable to meet the development needs of the new distribution network, and it is urgent to develop a fault section location algorithm suitable for distributed new energy grid connection. Summary of the invention

[0004] The purpose of this section is to summarize some aspects of embodiments of the present invention and briefly introduce some preferred embodiments. Some simplifications or omissions may be made in this section and the specification abstract and the invention title of this application to avoid blurring the purpose of this section, the specification abstract and the invention title, and such simplifications or omissions cannot be used to limit the scope of the present invention.

[0005] In view of the above existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method and system for locating a fault section in a distribution network, which can solve the problems mentioned in the background technology.

[0007] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0008] In a first aspect, the present invention provides a method for locating a fault section of a distribution network, comprising:

[0009] Establishing a target distribution network model, and performing a first preset section division based on the target distribution network model;

[0010] Acquire first target data of different short-circuit faults at different section positions after the first preset section is divided;

[0011] Performing a first preprocessing on the first target data to obtain a second target data set;

[0012] A first distribution network fault section location model is trained based on the second target data set, and the distribution network fault section location is performed according to the first distribution network fault section location model.

[0013] As a preferred solution of the method for locating a faulty section of a distribution network described in the present invention, the first preset section division includes: dividing the target distribution network into a number of different sections according to distance and load.

[0014] As a preferred solution of the method for locating a fault section of a distribution network according to the present invention, wherein: the first distribution network fault section locating model includes:

[0015] The first distribution network fault section location model is an arbitrary model built by a neural network algorithm, whose input is the second target data set and whose output is the distribution network fault section location probability.

[0016] As a preferred solution of the method for locating a fault section of a distribution network according to the present invention, the establishment of a target distribution network model comprises:

[0017] The target distribution network model is used to simulate different short-circuit faults at different section locations;

[0018] The target distribution network model includes at least a plurality of distributed power generation models, and the distributed power generation models include at least a wind power generation model and / or a photovoltaic power generation model.

[0019] As a preferred solution of the method for locating a fault section of a distribution network according to the present invention, the first preprocessing comprises:

[0020] Performing wavelet transform on the first target data to obtain third target data;

[0021] Performing a first normalization process on the third target data;

[0022] The third target data after the first normalization process is recorded as the second target data set.

[0023] As a preferred solution of the method for locating a fault section of a distribution network according to the present invention, wherein: recording the third target data after the first normalization process as the second target data set comprises:

[0024] Labeling the third target data after the first normalization process;

[0025] The content corresponding to each label is a multi-dimensional feature matrix;

[0026] All labeled data of the third target data after the first normalization processing are listed as a set, namely, the second target data set.

[0027] As a preferred solution of the distribution network fault section locating method described in the present invention, wherein: the first distribution network fault section locating model also includes: setting a five-layer network structure, and the five-layer network structure includes at least one dropout layer.

[0028] In a second aspect, the present invention provides a distribution network fault section location system, comprising:

[0029] A section division module, used for establishing a target distribution network model and performing a first preset section division based on the target distribution network model;

[0030] A data acquisition module, used for acquiring first target data of different short-circuit faults at different section positions after the first preset section is divided;

[0031] A data set acquisition module, used for performing a first preprocessing on the first target data to obtain a second target data set;

[0032] A positioning module is used to train a first distribution network fault section positioning model based on the second target data set, and locate the distribution network fault section according to the first distribution network fault section positioning model.

[0033] In a third aspect, the present invention provides a computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above-mentioned method when executing the computer program.

[0034] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program implements the steps of the method described above when executed by a processor.

[0035] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention proposes a distribution network fault section locating method and system, establishes a target distribution network model, and performs a first preset section division based on the target distribution network model; obtains first target data of different short-circuit faults at different section positions after the first preset section division; performs a first preprocessing on the first target data to obtain a second target data set; trains a first distribution network fault section locating model based on the second target data set, and locates the distribution network fault section according to the first distribution network fault section locating model. The present distribution network fault section location method builds an improved IEEE33 distribution network model including distributed power generation, and takes into account the nonlinear characteristics introduced to the distribution network system by distributed wind power generation and photovoltaic power generation; the present distribution network fault section location method uses wavelet packet transform to extract current signal characteristics, which can effectively capture the high-frequency characteristics of the signal; the artificial neural network of the present distribution network fault section location method can be well adapted to the aforementioned time-frequency feature matrix, and can achieve 99.8% fault section location accuracy on the constructed fault data set, and a dropout layer is set at the same time so that the algorithm can have good generalization ability; the present distribution network fault section location method will not be affected by the current flow direction changes and nonlinear characteristics brought about by the access of distributed power sources; the method can effectively cope with the challenges brought about by the grid connection of distributed new energy sources, and improve the accuracy and efficiency of fault location. The distribution network fault section location method and system of the present application provide strong technical support for the stable operation of the power system, help reduce power outage time, improve power supply reliability, and are of great significance to ensuring the stable operation of the social economy. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative labor. Among them:

[0037] Figure 1 A method flow chart of a method and system for locating a fault section in a distribution network provided by an embodiment of the present invention;

[0038] Figure 2 A topological diagram of an improved IEEE33 distribution network model including distributed generation, provided by a method and system for locating a fault section of a distribution network according to an embodiment of the present invention;

[0039] Figure 3 A grayscale image of a time-frequency feature matrix of a distribution network fault section location method and system provided by an embodiment of the present invention;

[0040] Figure 4A structural diagram of an artificial neural network ANN of a method and system for locating a fault section in a distribution network provided by an embodiment of the present invention;

[0041] Figure 5 A flow chart of a calculation example of a method and system for locating a fault section in a distribution network provided by an embodiment of the present invention;

[0042] Figure 6 An internal structural diagram of a computer device of a distribution network fault section locating method and system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0043] In order to make the above-mentioned purposes, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the drawings of the specification. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary persons in the art without creative work should fall within the scope of protection of the present invention.

[0044] Example 1

[0045] Reference Figure 1-Figure 6 , which is the first embodiment of the present invention, provides a method and system for locating a fault section of a distribution network, including:

[0046] There are some problems in the existing related technologies. For example, in terms of distribution network fault detection and location, traditional detection methods are often unable to accurately identify the fault section. Especially in the case of distributed power sources connected to the grid, the complexity and nonlinear characteristics of the current signal make fault location more difficult.

[0047] The present application provides a method that can effectively solve the above-mentioned problems. Next, multiple embodiments will be combined to explain in detail how to implement the method for locating the fault section of the distribution network;

[0048] Figure 1 A method flow chart of a method and system for locating a fault section of a distribution network is shown, including:

[0049] S101, establishing a target distribution network model, and performing a first preset section division based on the target distribution network model;

[0050] In an embodiment of the present application, the first preset section division includes: dividing the target distribution network into several different sections according to distance and load.

[0051] In an optional embodiment, the target distribution network is divided into several different sections according to distance and load. The distribution network can be monitored by distributed sensing. The distribution network is divided into different sections according to distance and load, and current sensors are installed at the head end of each section to measure the current signals IA, IB, IC of each phase of each section; voltage sensors can also be installed at the head end of each section to measure the voltage signals VA, VB, VC of each phase of each section. The data collected by these sensors can monitor the operating status of the distribution network in real time and provide basic data for fault detection and location.

[0052] In an optional embodiment, in addition to segmentation by installing various sensors, the distribution network can also be segmented by analyzing the topological structure and load distribution of the distribution network and using an intelligent algorithm to ensure that the fault characteristics in each segment have a certain similarity.

[0053] In an optional embodiment, the target distribution network is divided according to the distance and load. For example, a cluster analysis method can be used to group the nodes in the distribution network according to the similarity of distance and load to form several sections. The nodes in each section have similar current and voltage characteristics during normal operation. In this way, when a fault occurs, the section where the fault occurs can be quickly located by comparing the current and voltage data of each node. In addition, in order to improve the accuracy of fault detection, the operation history data of the distribution network can also be considered to dynamically adjust the section division to adapt to changes in the operation status of the distribution network.

[0054] In the embodiment of the present application, a current sensor is installed at the head end of each section to measure the current signal of each phase of each section.

[0055] In an embodiment of the present application, establishing a target distribution network model includes:

[0056] The target distribution network model is used to simulate different short-circuit faults at different section locations;

[0057] The target distribution network model includes at least a number of distributed power generation models, and the distributed power generation models include at least a wind power generation model and / or a photovoltaic power generation model.

[0058] In an alternative embodiment, the following is established: Figure 2 The distribution network model shown is an improved IEEE33 distribution network model including distributed generation built using PSCAD / EMTDC. Different types of faults at 7 locations are configured in the built PSCAD / EMTDC model, and models including wind power generation and photovoltaic power generation are added.

[0059] In an optional embodiment, it is also possible to use Figure 2The main model is shown, but different types of faults are selected in more or fewer locations, and relevant technicians can design it according to actual needs.

[0060] It should be noted that the model including wind power generation and photovoltaic power generation can provide a more accurate simulation environment for fault detection and location. By simulating different types of short-circuit faults, rich fault data can be collected to provide a basis for subsequent fault feature extraction and analysis. In addition, the model can also simulate the dynamic characteristics of distributed power sources, such as the randomness of wind power generation and the intermittent nature of photovoltaic power generation, so as to more realistically reflect the operating status of the actual distribution network under the condition of new energy grid connection. The distribution network model includes distributed generation, so that the fault current contains nonlinear characteristics, making the fault model closer to the actual working conditions.

[0061] S102, obtaining first target data of different short-circuit faults at different section positions after the first preset section is divided;

[0062] In the embodiment of the present application, data of different short-circuit faults at different section positions after the first preset section division is obtained and first fault labeling is performed;

[0063] In an optional embodiment, the first fault labeling is to label the collected first time domain signal data with a corresponding fault type, for example, labels such as "fault type A" and "fault type B" may be used to facilitate subsequent data classification and analysis. Other labeling methods may also be used, for example, digital coding or color coding to distinguish different fault types.

[0064] In an embodiment of the present application, labels "AG", "BG", "CG", "AB", "AC", "BC", "AB-G", "AC-G", "BC-G", "ABC", and "ABC-G" are used to identify different types of short-circuit faults that may occur in a distribution network, wherein AG represents a single-phase short-circuit grounding fault of phase A (single-phase fault, the fault point is between phase A and ground), BG represents a single-phase short-circuit grounding fault of phase B (single-phase fault, the fault point is between phase B and ground), CG represents a single-phase short-circuit grounding fault of phase C (single-phase fault, the fault point is between phase C and ground), AB represents a two-phase short-circuit fault of phase A and phase B (two-phase fault, the fault point is between phase A and phase B), and AC represents a two-phase short-circuit fault of phase A and phase C (two-phase fault). Fault, the fault point is between phase A and phase C), BC represents a two-phase short-circuit fault between phase B and phase C (two-phase fault, the fault point is between phase B and phase C), AB-G represents a two-phase short-circuit grounding fault between phase A and phase B (two-phase grounding fault, the fault point is between phase A, phase B and ground), AC-G represents a two-phase short-circuit grounding fault between phase A and phase C (two-phase grounding fault, the fault point is between phase A, phase C and ground), BC-G represents a two-phase short-circuit grounding fault between phase B and phase C (two-phase grounding fault, the fault point is between phase B, phase C and ground), ABC represents a three-phase short-circuit fault (three-phase fault, the fault point is between phase A, phase B and phase C), ABC-G represents a three-phase short-circuit grounding fault (three-phase grounding fault, the fault point is between phase A, phase B, phase C and ground). The reason for the design of this application is to combine it with the subsequent three-phase current characteristics so as to achieve more accurate fault type identification in the fault classification model. In this way, it is ensured that the model has high adaptability and accuracy when facing real power grid faults.

[0065] It should be noted that obtaining the first target data of different short-circuit faults at different section locations after the first preset section division can provide a more accurate reference basis for fault detection and positioning of the distribution network. By simulating short-circuit faults in different sections, rich fault data can be obtained. These data can not only include time domain signals of current and voltage, but also cover environmental parameters when the fault occurs, such as temperature, humidity, etc. The collection and analysis of these data are crucial to establishing an accurate fault feature database, which helps to improve the accuracy and response speed of fault diagnosis. In addition, by simulating different types of short-circuit faults, the accuracy and reliability of the distribution network model can be verified to ensure that faults can be effectively identified and located in practical applications, thereby ensuring the safe and stable operation of the distribution network.

[0066] S103, performing a first preprocessing on the first target data to obtain a second target data set;

[0067] In the embodiment of the present application, the first preprocessing includes:

[0068] Performing wavelet transform on the first target data to obtain third target data;

[0069] performing a first normalization process on the third target data;

[0070] The third target data after the first normalization process is recorded as the second target data set.

[0071] In the embodiments of the present application, wavelet transform is a common signal processing method that can effectively separate different frequency components of a signal and retain signal characteristics in both the time domain and the frequency domain. However, wavelet transform has certain defects. It can only effectively process the low-frequency components of the signal and has limited analysis of high-frequency components. Wavelet packet transform is developed on the basis of wavelet transform and can solve the defect that high frequencies of wavelet transform are difficult to analyze. Wavelet packet transform can effectively extract the characteristics of distribution network fault signals in different frequency bands. The amplitude of transient signals of distribution network faults is larger than that of steady-state signals and contains rich fault information. Wavelet packet transform is used to extract the characteristics of current signals collected by sensor nodes.

[0072] Mathematically, a wavelet packet consists of a set of linear combinations of wavelet functions, which are generated by the following recursive relationship:

[0073]

[0074] Among them, φ 0 (t) is the calibration function, φ 1 (t) is the wavelet function. h(n) and g(n) represent the coefficients of a pair of orthogonal mirror filters associated with the calibration function and the wavelet function. In addition, h(n) and g(n) are related by g(n) = (-1) n h(1-n) are related to each other. For each step of decomposition, the input discrete signal is decomposed into a low-frequency coarse approximation signal and a high-frequency detail part. The time domain signal x(t) can be decomposed into

[0075]

[0076] Therefore, the signal x(t) can be expressed as:

[0077]

[0078] Where j and k represent the number of decomposition levels and sub-bands, respectively.

[0079] Furthermore, the current signal features extracted by multiple sensor nodes are arranged to construct a multi-dimensional feature matrix, and the processed current signals include three-phase current and zero-sequence current. Wavelet packet transform is used to extract the time-frequency features of the fault three-phase current and zero-sequence current, and normalization is performed.

[0080]

[0081] In the embodiment of the present application, recording the third target data after the first normalization process as the second target data set includes:

[0082] Labeling the third target data after the first normalization process;

[0083] The content corresponding to each label is a multi-dimensional feature matrix;

[0084] All labeled data of the third target data after the first normalization processing are listed as a set, namely, the second target data set.

[0085] In an optional embodiment, labeling the third target data after the first normalization processing can be achieved through the following steps: first, assigning a unique label to each fault type according to the fault type and characteristics; second, associating the characteristic data of each fault instance with the corresponding label to form a labeled data pair; finally, integrating these data pairs into a complete data set, namely the second target data set.

[0086] It should be noted that in this way, it can be ensured that in the subsequent fault classification model training process, each data point carries clear fault type information, thereby improving the classification accuracy of the model. In addition, labeling processing also helps to quickly identify and locate faults after model training, because the model can learn the mapping relationship between different fault types and features. Ultimately, through this systematic and standardized data processing process, the overall performance of the distribution network fault detection and location system can be effectively improved.

[0087] In an optional embodiment, the specific labeling method may be a coding method based on fault characteristics. For example, binary coding may be used to represent different fault types, where each binary bit represents a specific fault characteristic. For example, one binary bit may represent whether there is a single-phase ground fault, another bit may represent whether there is a two-phase short circuit fault, and so on.

[0088] In an optional embodiment, octal or hexadecimal representation can also be used, for example, the encoding method of the fault feature is combined with the specific parameters of the fault type to form a comprehensive fault feature encoding table. For example, for a single-phase grounding fault, the binary code "1000" can be used to represent it, while for a two-phase short circuit fault, "0100" can be used. In this way, complex fault features can be converted into a digital form that is easy for a computer to process, thereby simplifying the fault diagnosis process. In addition, this encoding method can also be extended to more fault types and features, improving the flexibility and scalability of the system. In practical applications, this encoding table can be used as part of a fault diagnosis system to help quickly identify and locate faults in a distribution network. Ultimately, through this precise fault feature extraction and encoding method, the distribution network fault section positioning system can achieve efficient and accurate fault detection and positioning, ensuring the stable operation of the power system.

[0089] In an embodiment of the present application, a two-dimensional reconstruction is used to label the third target data after the first normalization processing. In the two-dimensional reconstruction process, the time-frequency characteristics of the fault signal are mapped to a two-dimensional plane to form a reconstruction pair. By analyzing the distribution pattern of the reconstruction pair, the characteristic patterns corresponding to different fault types can be identified. The reconstruction pair is associated with the corresponding fault type label to form a labeled data pair. Finally, all the labeled data pairs are integrated into a complete data set, namely the second target data set. Through the processing of the above-mentioned two-dimensional reconstruction pairs, the distribution network fault section positioning system can more accurately identify and locate faults, thereby improving the reliability and safety of the entire power system.

[0090] It should be noted that the difference between using two-dimensional reconstruction and fault feature coding for this application is that the processing method of two-dimensional reconstruction can more intuitively display the characteristics of the fault signal, while fault feature coding provides a more compact and standardized way to represent fault information. In practical applications, these two methods can complement each other to achieve the best fault detection and location effect. For example, in some cases, two-dimensional reconstruction may be more suitable for intuitively displaying and analyzing the time-frequency characteristics of the fault signal, while fault feature coding may be more effective when rapid processing and transmission of fault information is required.

[0091] S104: training a first distribution network fault section location model based on the second target data set, and locating the distribution network fault section according to the first distribution network fault section location model.

[0092] In an optional embodiment, the first distribution network fault section location model can be designed by a neural network algorithm (specifically, a convolutional neural network (CNN) or a recurrent neural network (RNN)), and the first distribution network fault section location model can also be established by methods other than the neural network algorithm, for example, support vector machine (SVM), decision tree, random forest and other machine learning algorithms. However, no matter what method is used to train the first distribution network fault section location model, as long as the input of the model is the second target data set, and the output is the distribution network fault section location probability or any model related to the distribution network fault section location should be within the protection scope of this application;

[0093] In an embodiment of the present application, the first distribution network fault section location model includes: the first distribution network fault section location model is an arbitrary model built by a neural network algorithm, whose input is the second target data set and whose output is the distribution network fault section location probability.

[0094] In the embodiment of the present application, an artificial neural network is selected to design the first distribution network fault section location model, and the structure of the artificial neural network is designed, and the network structure is adapted to the multidimensional feature matrix. Compared with traditional machine learning, the artificial neural network has the ability of self-adaptation, self-organization and self-learning. It can fully approximate any complex linear relationship.

[0095] In the embodiment of the present application, the multi-scale characteristics of feature extraction by wavelet packet transform are taken into consideration, convolutional layers, pooling layers, hidden layers, etc. are set in the network, the network parameters are reasonably adjusted, and a dropout layer is set to enhance the generalization ability of the network. Figure 4 This is the structure diagram of the artificial neural network ANN. CONV is the convolution layer of the network, which changes the dimension of the matrix by moving the convolution kernel to extract features. Maxpool is the pooling layer, which is sandwiched between the convolution layers and is used to compress data and reduce the number of parameters to reduce the degree of overfitting. The Dropout layer achieves the effect of enhancing the generalization ability of the network by covering part of the data. The results of the previous calculation are converted into probabilities through the activation function, and the result with the highest probability is the segment positioning result.

[0096]

[0097] Furthermore, the designed artificial neural network ANN is used to train the multi-dimensional feature matrix to continuously optimize the model parameters and improve the model accuracy.

[0098] In an embodiment of the present application, the first distribution network fault section location model also includes: setting a five-layer network structure, the five-layer network structure includes at least one dropout layer.

[0099] Figure 3It is a grayscale image of fault features at different locations in the fault dataset. The numbers in the image indicate the section to which the fault belongs. 0 means that the distribution network is operating normally and there is no fault.

[0100] In the embodiment of the present application, PSCAD / EMTDC is used to build an IEEE33 distribution network model. The model parameters and topology structure match the IEEE33 node distribution network model. Models of wind power generation and photovoltaic power generation are designed, and distributed wind power generation and photovoltaic power generation are added to some nodes in the distribution network model. And sensor nodes are added to monitor the current. 10 different short-circuit faults (AG, BG, CG, AB, AC, BC, AB-G, AC-G, BC-G, ABC-G) are set in 7 sections of the model. Faults with different phases and different transition resistances are set. The model is simulated, and the sensor nodes collect fault currents.

[0101] The current data is imported into MATLAB, and the wavelet packet transform function is used to perform wavelet packet transform on the current to obtain the high-frequency features of the current signal. The high-frequency features of the current signal extracted by multiple sensor nodes are arranged to construct a multidimensional feature matrix. It can be seen that when faults occur in different sections, the feature matrix is ​​different. The multidimensional feature matrices extracted from different short-circuit faults in different section locations are labeled to form a fault data set.

[0102] In python, the artificial neural network ANN is built using the pytorch framework. The structure and feature matrix of the artificial neural network are adapted. The artificial neural network is trained and the parameters are optimized, which can achieve an accuracy of more than 99% on both the training set and the test set. The wavelet packet transform and the artificial neural network ANN are combined to construct a distribution network fault section location algorithm.

[0103] In the embodiment of the present application, the structure diagram of the specific artificial neural network ANN is as follows: Figure 4 As shown in the figure, SENet (Squeeze-and-Excitation Network): The input is a two-dimensional matrix representing a specific input data. SENet is a module for enhancing the correlation between channels, but its internal details are not expanded in this simplified illustration.

[0104] CONV 3x3, stride=1 represents the first layer of convolution operation, using a convolution kernel of size 3x3 and a stride of 1. This means that the convolution kernel moves one step at a time, covering the entire input matrix. After the convolution operation, a new three-dimensional tensor is obtained, which represents the filtered feature map.

[0105] Max POOL 2x2, stride = 2 represents the maximum pooling operation, using a window of size 2x2 and a stride of 2. This reduces the spatial size of the feature map, helps reduce computational complexity and prevent overfitting.

[0106] The output is a smaller three-dimensional tensor where each element is the result of the maximum value in the corresponding region of the original feature map.

[0107] CONV 3x3, stride=1 indicates the second convolution operation, which also uses a 3x3 convolution kernel with a stride of 1. Further feature extraction is performed.

[0108] Dropout is a regularization technique that randomly drops a portion of nodes during training to reduce the risk of overfitting.

[0109] CONV 3x3, stride=1 indicates the third layer of convolution operation, which continues to use a 3x3 convolution kernel with a stride of 1. This series of convolution operations aims to extract increasingly abstract features from the original input.

[0110] Flatten means flattening the output of the last convolutional layer into a one-dimensional array so that it can be connected to the fully connected layer.

[0111] Fully Connected Layer (FC) means that the fully connected layer receives the flattened feature vector and maps it to the dimension corresponding to the number of categories. Here, a fully connected layer with multiple nodes is shown.

[0112] The Softmax function is applied to the output of the fully connected layer to produce a probability distribution representing the likelihood of each category. In this example, the final output is a 7-dimensional probability vector, which means that the model is used to solve 7-category classification problems.

[0113] In summary, the present invention proposes a method for locating a distribution network fault section, establishes a target distribution network model, and performs a first preset section division based on the target distribution network model; obtains first target data of different short-circuit faults at different section positions after the first preset section division; performs a first preprocessing on the first target data to obtain a second target data set; trains a first distribution network fault section locating model based on the second target data set, and locates the distribution network fault section according to the first distribution network fault section locating model. The present distribution network fault section location method builds an improved IEEE33 distribution network model including distributed power generation, and takes into account the nonlinear characteristics introduced to the distribution network system by distributed wind power generation and photovoltaic power generation; the present distribution network fault section location method uses wavelet packet transform to extract current signal characteristics, which can effectively capture the high-frequency characteristics of the signal; the artificial neural network of the present distribution network fault section location method can be well adapted to the aforementioned time-frequency feature matrix, and can achieve 99.8% fault section location accuracy on the constructed fault data set, and a dropout layer is set at the same time so that the algorithm can have good generalization ability; the present distribution network fault section location method will not be affected by the current flow direction changes and nonlinear characteristics brought about by the access of distributed power sources; the method can effectively cope with the challenges brought about by the grid connection of distributed new energy sources, and improve the accuracy and efficiency of fault location. The distribution network fault section location method and system of the present application provide strong technical support for the stable operation of the power system, help reduce power outage time, improve power supply reliability, and are of great significance to ensuring the stable operation of the social economy.

[0114] This embodiment also provides a distribution network fault section locating system, including:

[0115] A section division module, used for establishing a target distribution network model and performing a first preset section division based on the target distribution network model;

[0116] A data acquisition module, used for acquiring first target data of different short-circuit faults at different section positions after the first preset section is divided;

[0117] A data set acquisition module, used for performing a first preprocessing on the first target data to obtain a second target data set;

[0118] The positioning module is used to train the first distribution network fault section positioning model based on the second target data set, and locate the distribution network fault section according to the first distribution network fault section positioning model.

[0119] The above-mentioned unit modules may be embedded in or independent of a processor in a computer device in the form of hardware, or may be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to the above-mentioned modules.

[0120] This embodiment also provides a computer device, which may be a terminal, and its internal structure diagram may be as shown in FIG. Figure 6 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a method for locating a fault section of a distribution network is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0121] This embodiment further provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0122] Establishing a target distribution network model, and performing a first preset section division based on the target distribution network model;

[0123] Acquire first target data of different short-circuit faults at different section positions after the first preset section is divided;

[0124] Performing a first preprocessing on the first target data to obtain a second target data set;

[0125] A first distribution network fault section location model is trained based on the second target data set, and the distribution network fault section location is performed according to the first distribution network fault section location model.

[0126] Example 2

[0127] Reference Figure 5 , which is an embodiment of the present invention, provides a method and system for locating a fault section in a distribution network. In order to verify the beneficial effects of the present invention, scientific demonstration is carried out through experiments.

[0128] Use PSCAD / EMTDC to build an IEEE33 distribution network model. The model parameters and topology match the IEEE33 node distribution network model. Design models for wind power generation and photovoltaic power generation, and add distributed wind power generation and photovoltaic power generation to some nodes in the distribution network model. Add sensor nodes to monitor the current. Set 10 different short-circuit faults (AG, BG, CG, AB, AC, BC, AB-G, AC-G, BC-G, ABC) in 7 sections of the model. Set faults with different phases and different transition resistances. Simulate the model, and the sensor nodes collect the fault current.

[0129] The current data is imported into MATLAB, and the wavelet packet transform function is used to perform wavelet packet transform on the current to obtain the high-frequency features of the current signal. The high-frequency features of the current signal extracted by multiple sensor nodes are arranged to construct a multidimensional feature matrix. It can be seen that when faults occur in different sections, the feature matrix is ​​different. The multidimensional feature matrices extracted from different short-circuit faults in different section locations are labeled to form a fault data set.

[0130] In python, the artificial neural network ANN is built using the pytorch framework. The structure and feature matrix of the artificial neural network are adapted. The artificial neural network is trained and the parameters are optimized, which can achieve an accuracy of more than 99% on both the training set and the test set. The wavelet packet transform and the artificial neural network ANN are combined to construct a distribution network fault section location algorithm.

[0131] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

[0132] Those skilled in the art will appreciate that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application can adopt the form of complete hardware embodiments, complete software embodiments, or embodiments in combination with software and hardware. Moreover, the present application can adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that contain computer-usable program code. The scheme in the embodiments of the present application can be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0133] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0134] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0135] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0136] Although the preferred embodiments of the present application have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications falling within the scope of the present application.

[0137] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is also intended to include these modifications and variations.

Claims

1. A method for locating a fault section of a distribution network, characterized in that: include: Establishing a target distribution network model, and performing a first preset section division based on the target distribution network model; Acquire first target data of different short-circuit faults at different section positions after the first preset section is divided; Performing a first preprocessing on the first target data to obtain a second target data set; A first distribution network fault section location model is trained based on the second target data set, and the distribution network fault section location is performed according to the first distribution network fault section location model.

2. The method for locating a fault section of a distribution network according to claim 1, characterized in that: The first preset section division includes: dividing the target distribution network into a plurality of different sections according to distance and load.

3. The method for locating a fault section of a distribution network according to claim 2, characterized in that: The first distribution network fault section location model includes: The first distribution network fault section location model is an arbitrary model built by a neural network algorithm, whose input is the second target data set and whose output is the distribution network fault section location probability.

4. The method for locating a fault section of a distribution network according to claim 3, characterized in that: The establishing of the target distribution network model comprises: The target distribution network model is used to simulate different short-circuit faults at different section locations; The target distribution network model includes at least a plurality of distributed power generation models, and the distributed power generation models include at least a wind power generation model and / or a photovoltaic power generation model.

5. The method for locating a fault section of a distribution network according to claim 4, characterized in that: The first preprocessing comprises: Performing wavelet transform on the first target data to obtain third target data; Performing a first normalization process on the third target data; The third target data after the first normalization process is recorded as the second target data set.

6. The method for locating a fault section of a distribution network according to claim 5, characterized in that: The step of recording the third target data after the first normalization process as the second target data set includes: Labeling the third target data after the first normalization process; The content corresponding to each label is a multi-dimensional feature matrix; All labeled data of the third target data after the first normalization processing are listed as a set, namely, the second target data set.

7. The method for locating a fault section of a distribution network according to claim 6, characterized in that: The first distribution network fault section location model also includes: setting a five-layer network structure, and the five-layer network structure includes at least one dropout layer.

8. A distribution network fault section location system, characterized in that: include: A section division module, used for establishing a target distribution network model and performing a first preset section division based on the target distribution network model; A data acquisition module, used for acquiring first target data of different short-circuit faults at different section positions after the first preset section is divided; A data set acquisition module, used for performing a first preprocessing on the first target data to obtain a second target data set; A positioning module is used to train a first distribution network fault section positioning model based on the second target data set, and locate the distribution network fault section according to the first distribution network fault section positioning model.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to 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 according to any one of claims 1 to 7 are implemented.

Images