Low-resistance grounding system latent grounding fault detection method and system

Abstract

The invention discloses a method and a system for detecting a latent grounding fault of a small-resistance grounding system, relates to the technical field of fault detection, and solves the problem that the single-phase latent grounding fault of a power distribution network is difficult to detect in the prior art. According to the method, neutral line zero-sequence current waveforms corresponding to events, such as single-phase latent ground faults and system asymmetry events, which can cause increase of the amplitude of neutral line zero-sequence current under different conditions are obtained by establishing a single-phase latent ground fault simulation model of the small-resistance grounding system; intercepting a waveform near a position where the amplitude of the zero-sequence current of the neutral line exceeds a setting value, and converting the waveform into a digital signal form through a signal sampling module; converting the digital signal into a two-dimensional graph through a Grubm angle field; and after the two-dimensional graphs are labeled, inputting the two-dimensional graphs as training data into the convolutional neural network for training and testing, and classifying the two-dimensional graphs corresponding to different events through the convolutional neural network to realize the distinguishing between the single-phase latent ground fault and the system asymmetric event.

Description

Technical Field

[0001] The present invention relates to the technical field of fault detection, and in particular to a method and system for detecting latent grounding faults in a low-resistance grounding system. Background Art

[0002] With the rapid advancement of China's urbanization process, the cabling rate of urban medium-voltage distribution networks has increased significantly. This change has brought many advantages, such as higher power supply reliability and smaller space occupancy. However, the resulting increase in capacitive current has also posed a challenge to traditional power system protection methods. Specifically, in urban distribution networks with a high degree of cabling, due to the capacitive effect of the cable itself, the capacitive current has increased significantly compared to traditional overhead lines, which makes the traditional arc suppression coil compensation method incapable of coping with large capacitive currents, making it difficult to effectively deal with ground faults caused by high capacitive currents.

[0003] In order to overcome the above difficulties, the low-resistance grounding system is widely used in urban distribution networks as an effective solution. A small resistor is connected between the neutral point of the transformer and the ground to achieve the purpose of fast fault removal and limit the overvoltage level. Compared with the traditional arc suppression coil grounding method, the low-resistance grounding system has the advantages of fast response speed and low overvoltage level, and can better meet the needs of modern urban distribution networks. However, the current low-resistance grounding system mainly relies on time-limited zero-sequence overcurrent protection as the main protection means. This protection mechanism sets the threshold according to the maximum capacitive current, and the general setting range is 40A to 60A. However, according to actual application experience, such a setting value can only ensure the detection of faults with a grounding resistance not exceeding 85Ω to 140Ω. For those hidden or latent grounding faults with a grounding resistance greater than 140Ω, the zero-sequence current amplitude generated may be lower than the protection action threshold, so the protection action cannot be triggered, which may cause the fault to be not discovered in time. The existence of this situation seriously affects the operational stability of the power system and poses a potential threat to public safety. Failure to identify and handle these latent grounding faults in a timely manner may lead to the deterioration of the fault situation, ultimately causing serious economic losses or even endangering life safety.

[0004] In view of this, it is particularly important to study and develop a more sensitive and reliable detection method to identify latent ground faults in low-resistance grounding systems. Summary of the 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 detecting latent grounding faults in a low-resistance grounding system to solve the problem that the time-limited zero-sequence overcurrent protection relied on by the traditional low-resistance grounding system is not sensitive to the detection of latent grounding faults with a grounding resistance greater than 140Ω, which may result in the fault not being discovered in time, thereby threatening the stable operation of the power system and public safety.

[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 detecting latent grounding faults in a low-resistance grounding system, comprising: establishing a single-phase latent grounding fault simulation model for a low-resistance grounding system, and acquiring a first current waveform; detecting an existing latent grounding fault, and converting the intercepted second current waveform into a two-dimensional graph; inputting the two-dimensional graph into a first neural network for training to classify the two-dimensional graphs corresponding to different events, thereby distinguishing between single-phase latent grounding faults and system asymmetric events.

[0009] As a preferred solution of the method for detecting latent grounding faults in a low-resistance grounding system according to the present invention, the detecting of the latent grounding faults includes:

[0010] Set the first current amplitude setting value;

[0011] When the amplitude of the zero-sequence current of the system neutral line is greater than the first current amplitude setting value, it is determined that at least a single-phase latent grounding fault or a system asymmetry event occurs in the system.

[0012] As a preferred solution of the method for detecting latent ground faults in a low-resistance grounding system according to the present invention, the intercepted second current waveform is a zero-sequence current waveform within a first time threshold at an amplitude exceeding the limit.

[0013] As a preferred solution of the method for detecting latent ground faults in a low-resistance grounding system according to the present invention, the step of converting the intercepted second current waveform into a two-dimensional graph comprises:

[0014] Sampling the analog signal at a second time threshold to generate discrete sample points;

[0015] quantizing the amplitude of the sampled discrete signal, converting it into a finite discrete value, and converting the discrete value into a binary form of a digital signal;

[0016] Convert to 2D graphics via Gram's angle field.

[0017] As a preferred solution of the method for detecting latent grounding faults in a low-resistance grounding system of the present invention, wherein: the first neural network includes an input layer, a convolution layer, an activation function layer, a pooling layer, a fully connected layer and an output layer;

[0018] The input layer receives a two-dimensional graph generated by Gram's angular field transformation, the convolution layer extracts features from the input data through a convolution kernel, the pooling layer is used to reduce the data dimension, the fully connected layer maps the features extracted by the convolution layer to an output space of a fixed dimension for classification or regression, and the output layer outputs a probability distribution of the two-dimensional graph classification corresponding to single-phase latent grounding faults and system asymmetry events.

[0019] As a preferred solution of the method for detecting latent grounding faults in a low-resistance grounding system of the present invention, the single-phase latent grounding fault simulation model of the low-resistance grounding system includes:

[0020] Neutral grounding resistance and feeder lengths of varying lengths;

[0021] In the simulation model, the values ​​of the neutral point grounding resistance, feeders of different lengths, and the positive-sequence parameters and zero-sequence parameters of the line are set respectively.

[0022] As a preferred solution of the method for detecting latent ground faults in a low-resistance grounding system of the present invention, the conversion of the Gram angle field into a two-dimensional graph comprises:

[0023] Collecting and preprocessing the amplitude data of the neutral line zero-sequence current, and representing the amplitude of the preprocessed neutral line zero-sequence current and its corresponding timestamp by the angle and radius of polar coordinates respectively;

[0024] Divide the interval into several parts, associate the remaining points with the time series data by discarding 0, and make each point fall into a unit circle in polar coordinates;

[0025] The GAF algorithm is used to reconstruct the neutral line zero-sequence current time series, and two different images are generated based on different formulas, namely the Gram angle sum field or the Gram angle difference field.

[0026] In a second aspect, the present invention provides a system for detecting latent ground faults in a low-resistance grounding system, comprising:

[0027] A fault simulation module, used to establish a single-phase latent ground fault simulation model of a low-resistance grounding system and obtain a first current waveform;

[0028] A fault detection and waveform conversion module, used to detect the existence of a latent ground fault and convert the intercepted second current waveform into a two-dimensional graph;

[0029] The fault classification module is used to input the two-dimensional graph into the first neural network for training, so as to classify the two-dimensional graphs of different events and distinguish between single-phase latent grounding faults and system asymmetric events.

[0030] In a third aspect, the present invention provides an electronic device, comprising:

[0031] Memory and processor;

[0032] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method for detecting latent grounding faults in a low-resistance grounding system are implemented.

[0033] In a fourth aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for detecting latent grounding faults in a low-resistance grounding system.

[0034] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention provides a method and system for detecting latent grounding faults in a low-resistance grounding system, by establishing a simulation model of single-phase latent grounding faults in a low-resistance grounding system, and using a convolutional neural network to intelligently classify the neutral line zero-sequence current waveform converted into a two-dimensional graph, thereby achieving effective distinction between single-phase latent grounding faults and system asymmetric events, and can not only accurately capture and identify high grounding resistance faults that are difficult to detect by traditional protection mechanisms, and improve the speed and accuracy of fault diagnosis, but also reduce the false alarm rate, enhance the operating stability and safety of the power system, and provide strong technical support for timely troubleshooting, ensuring power grid reliability, and maintaining public safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0036] Figure 1 A schematic diagram of the overall process logic of a method for detecting a latent ground fault in a low-resistance grounding system according to an embodiment of the present invention;

[0037] Figure 2 A schematic diagram of an analog-to-digital conversion module of a method for detecting a latent ground fault in a low-resistance grounding system according to an embodiment of the present invention;

[0038] Figure 3A schematic diagram of an analog-to-digital conversion module constructed by MATLAB / Simulink for a latent grounding fault detection method for a low-resistance grounding system according to an embodiment of the present invention;

[0039] Figure 4 A schematic diagram of a Gram angle field conversion process of a method for detecting latent ground faults in a low-resistance grounding system according to an embodiment of the present invention;

[0040] Figure 5 A schematic diagram of a single-phase latent ground fault simulation model of a low-resistance grounding system constructed by MATLAB / Simulink according to a latent ground fault detection method of a low-resistance grounding system according to an embodiment of the present invention;

[0041] Figure 6 A schematic diagram of a neutral line zero-sequence current waveform when a B-phase 3000 ohm single-phase arc grounding fault occurs in line L3 obtained by running a simulation model of a latent grounding fault detection method for a low-resistance grounding system according to an embodiment of the present invention;

[0042] Figure 7 A schematic diagram of a zero-sequence current waveform of a neutral line when a load switching operation occurs in line L6 obtained by running a simulation model of a latent grounding fault detection method for a low-resistance grounding system according to an embodiment of the present invention;

[0043] Figure 8 A GASF graph obtained by using the Gram angle field in a method for detecting latent ground faults in a low-resistance grounding system according to an embodiment of the present invention;

[0044] Fig. 9 A GADF graph obtained by using the Gram angle field in a method for detecting latent ground faults in a low-resistance grounding system according to an embodiment of the present invention;

[0045] Fig.10 A schematic diagram of a training process of the accuracy rate obtained by running a convolutional neural network of a method for detecting latent grounding faults in a low-resistance grounding system according to an embodiment of the present invention;

[0046] Fig.11 A schematic diagram of a training process of losses obtained by running a convolutional neural network of a method for detecting latent grounding faults in a low-resistance grounding system according to an embodiment of the present invention;

[0047] Fig.12 A schematic diagram of training results obtained by running a convolutional neural network of a method for detecting latent grounding faults in a low-resistance grounding system according to an embodiment of the present invention;

[0048] Fig.13This is a schematic diagram of test results obtained by running a convolutional neural network of a method for detecting latent grounding faults in a low-resistance grounding system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0049] 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.

[0050] Example 1

[0051] Reference Figure 1 to Figure 4 As an embodiment of the present invention, a method for detecting latent ground faults in a low-resistance grounding system is provided. Figure 1 The specific steps shown include:

[0052] S100: establishing a single-phase latent ground fault simulation model for a low-resistance grounding system, and obtaining a first current waveform;

[0053] S200: detecting an existing latent ground fault and converting the intercepted second current waveform into a two-dimensional graph;

[0054] S300: Inputting the converted two-dimensional graph into a first neural network for training to classify the two-dimensional graphs corresponding to different events, thereby distinguishing between single-phase latent grounding faults and system asymmetry events.

[0055] It should be noted that the present invention provides a method and system for detecting latent grounding faults in a low-resistance grounding system. By establishing a simulation model of single-phase latent grounding faults in a low-resistance grounding system and using a convolutional neural network to intelligently classify the neutral line zero-sequence current waveform converted into a two-dimensional graph, the effective distinction between single-phase latent grounding faults and system asymmetric events is achieved. It can not only accurately capture and identify high grounding resistance faults that are difficult to detect by traditional protection mechanisms, improve the speed and accuracy of fault diagnosis, but also reduce the false alarm rate, enhance the operating stability and safety of the power system, and provide strong technical support for timely troubleshooting, ensuring power grid reliability and maintaining public safety.

[0056] In the embodiment of the present application, the above step S100 includes the following sub-steps A1-A2;

[0057] In A1: establish a single-phase latent ground fault simulation model for a low-resistance grounding system;

[0058] Among them, the single-phase latent ground fault simulation model of the low-resistance grounding system includes the neutral point grounding resistance and six feeders of different lengths. In the simulation model, the values ​​of the neutral point grounding resistance, six feeders of different lengths, and the positive-sequence parameters and zero-sequence parameters of the line are set respectively.

[0059] In A2: obtaining a first current waveform through a single-phase latent grounding fault simulation model of a low-resistance grounding system;

[0060] In an optional embodiment, the first current waveform may be a fault instantaneous current waveform, that is, when a single-phase latent ground fault occurs, the current suddenly increases to form a peak or step-change waveform; the first current waveform may be a steady-state fault current waveform, that is, a current waveform within a period of time after the fault occurs; the first current waveform may also be a transition process current waveform or a non-periodic component current waveform;

[0061] In the embodiment of the present application, the first current waveform obtained is a neutral line zero-sequence current waveform corresponding to events that may cause an increase in the amplitude of the neutral line zero-sequence current, including single-phase latent ground faults and system asymmetry events.

[0062] Specifically, through the single-phase latent grounding fault simulation model of the low-resistance grounding system, the single-phase latent grounding faults under different fault lines, different fault times, and different transition resistances and the events that may cause the neutral line zero-sequence current amplitude to increase are simulated to obtain the neutral line zero-sequence current waveform.

[0063] It should be noted that the above step S100 achieves a high degree of simulation of the real power system environment, can accurately reflect the characteristics of latent ground faults of different types and locations, and provide a valuable data basis for subsequent fault detection and classification. It can not only improve the sensitivity and accuracy of fault detection, but also enhance the operational stability and safety of the power system, and ensure faster and more reliable fault response.

[0064] In the embodiment of the present application, the above step S200 includes the following sub-steps B1-B3;

[0065] In B1: detecting the existence of a latent ground fault by setting the first current amplitude setting value;

[0066] In an optional embodiment, the first current amplitude setting value is a threshold value for determining whether there is a latent ground fault, which is usually preset according to system characteristics and protection requirements. The first current amplitude setting value can be a specific value obtained by considering historical data and experience, system parameter calculation, the current corresponding to the minimum detectable ground resistance, and the noise and interference level;

[0067] In the embodiment of the present application, the first current amplitude setting value is set to 1.5A.

[0068] Specifically, when a single-phase grounding fault occurs in a low-resistance grounding system or an asymmetric event occurs in the system (such as capacitor switching, load switching), the neutral line zero-sequence voltage amplitude will increase. This embodiment uses the neutral line zero-sequence current amplitude as a detection criterion for latent grounding faults in low-resistance grounding systems. When the system neutral line zero-sequence current amplitude is greater than the first current amplitude setting value, it is determined that the system has at least a single-phase latent grounding fault or a system asymmetric event.

[0069] Specifically, the setting of the neutral line zero-sequence current amplitude is as follows: According to the general parameters of engineering experience, the maximum unbalanced zero-sequence current corresponding to a single-phase grounding fault in a pure overhead line and a pure cable line is about 0.37A and 0.26A, respectively. Since the protection device receives a three-fold zero-sequence current electrical signal, the neutral line zero-sequence current setting value should be greater than 1.11A. Therefore, this embodiment sets the neutral line zero-sequence current amplitude setting value to 1.5A, which can detect a single-phase latent grounding fault of 3000 ohms.

[0070] In B2: intercepting the second current waveform;

[0071] In an optional embodiment, the second current waveform refers to a specific portion of the first current waveform that is cut out after a latent ground fault is detected, and this portion of the waveform contains key information about the fault, which may be: after a latent ground fault is detected, the system will focus on those current waveform segments that can reflect the fault characteristics, especially at the moment of the fault and in the short period of time thereafter; the second current waveform may be a transient response portion of the current sudden change when the fault occurs, or an abnormal current pattern that persists under steady-state fault conditions. In addition, it may also include current changes during the transition process, during which the current waveform often exhibits complex dynamic behaviors, such as oscillations or non-periodic fluctuations.

[0072] In the embodiment of the present application, the intercepted second current waveform is a zero-sequence current waveform at which the neutral line zero-sequence current amplitude exceeds the first current amplitude setting value;

[0073] Specifically, intercepting the zero-sequence current waveform of the neutral line where the zero-sequence current amplitude exceeds the first current amplitude setting value is specifically: intercepting the zero-sequence current waveform of the neutral line within the first time threshold where the amplitude exceeds the limit;

[0074] Specifically, the value of the first time threshold is within 0.005s before and 0.01s after the amplitude exceeds the limit.

[0075] In B3: converting the intercepted second current waveform into a two-dimensional graph;

[0076] Specifically, the sampling process of continuous signals is as follows: Figure 2 As shown, the analog-to-digital conversion module is Figure 3 As shown, it is composed of a sampler, a quantizer and an encoder, and the process of converting the second current waveform into a two-dimensional graph includes:

[0077] Zero-Order Hold (sampler): samples the analog signal at a second time threshold to generate discrete sample points;

[0078] Quantizer: quantizes the amplitude of the sampled discrete signal and converts it into a finite discrete value;

[0079] Data Type Conversion (Encoder): Converts discrete values ​​into binary form of digital signals;

[0080] Convert to two-dimensional graphics through Gram's angle field;

[0081] It should be noted that the second time threshold is the set sampling time, that is, 0.0001s.

[0082] Furthermore, the process of converting the Gram angle field into a two-dimensional graphic specifically includes:

[0083] The sensor collects the neutral line zero-sequence current amplitude data X = {x1, x2, x3, ..., xn-1, xn}, normalizes the collected data, and scales it to [-1, 1]. The formula is expressed as:

[0084]

[0085] The scaled neutral line zero-sequence current amplitude time series is converted into polar coordinates, and the amplitude of the neutral line zero-sequence current and its corresponding timestamp are represented by the angle and radius of the polar coordinates respectively;

[0086] Encode in polar coordinates, encode the scaled neutral line zero-sequence current amplitude with angle cosine, and encode the corresponding timestamp as radius. The angle and radius in polar coordinates are as follows:

[0087]

[0088] Among them, t i For x i The corresponding timestamp; M is the normalization factor;

[0089] The interval [0,1] is divided into several parts, and the remaining points are associated with the time series data after discarding 0, and each point is made to fall into a unit circle in polar coordinates; when time increases, the neutral line zero-sequence current time series will also be continuously distorted within this unit circle at different angles and radii;

[0090] The GAF algorithm is used to reconstruct the neutral line zero-sequence current time series. Two different images are generated based on different formulas, namely the Gram angle sum field or the Gram angle difference field. The expressions are as follows:

[0091]

[0092] Among them, A GASF (i,j) and A GADF (i, j) is the Gram matrix corresponding to the Gram angle sum field and the Gram angle difference field, θ i and θ j is the angle value of the time series data point in the polar coordinate system, and the subscripts i and j are used to represent the index of any two data points in the time series.

[0093] The complete Gram angle field conversion process is as follows: Figure 4 shown.

[0094] It should be noted that the above step S200 not only effectively captures the high ground resistance fault, but also ensures that the data best reflects the fault characteristics by focusing on the key current changes when the fault occurs. Converting the current waveform into a two-dimensional graph enhances the efficiency of data analysis, allowing the convolutional neural network to more effectively learn and distinguish single-phase latent ground faults from system asymmetry events.

[0095] In an optional embodiment, the first neural network may be a typical convolutional neural network architecture, including but not limited to the following configurations:

[0096] ① Classic convolutional neural network architecture: such as LeNet, AlexNet, VGG, etc.;

[0097] ② Improved convolutional neural network architecture: such as ResNet (residual network), DenseNet (densely connected network), etc.

[0098] ③Lightweight convolutional neural network architecture: such as MobileNet, SqueezeNet, etc.;

[0099] ④ Customized convolutional neural network architecture: According to the specific application scenario and data characteristics, a customized CNN specifically for latent ground fault detection in power systems can be designed, which may involve adjusting the size of the convolution kernel, adding an attention mechanism to emphasize important areas, or combining other types of layers (such as recurrent neural network RNN ​​layers) to process time series data;

[0100] In the embodiment of the present application, the first neural network adopts a convolutional neural network, including an input layer, a convolution layer, an activation function layer, a pooling layer, a fully connected layer and an output layer;

[0101] Specifically, the input layer receives a two-dimensional graph generated by converting the neutral line zero-sequence current waveform through the Gram angle field;

[0102] Specifically, the convolution layer extracts features from the input data through the convolution kernel (filter), and the convolution operation formula is expressed as:

[0103]

[0104] Among them, y(i,j) is the output value of the convolution operation, i represents the row index of the output result y(i,j), j represents the column index of the output result y(i,j), x is the input two-dimensional graphic data, x(i+m,j+n) represents the pixel value of the input two-dimensional graphic at the position (i+m,j+n), ω is the convolution kernel weight, ω(m,n) is the weight matrix of the convolution kernel, m is the row index of the convolution kernel ω(m,n), n is the column index of the convolution kernel ω(m,n), the position is (m,n) which is used to loop through each weight in the convolution kernel, b is the bias value; the convolution kernel ω(m,n) slides on the input two-dimensional graphic x(i+m,j+n) (from left to right, from top to bottom), and at each sliding position, the corresponding area of ​​the input feature map is multiplied point by point by the convolution kernel, the sum is calculated, and the bias b is added to obtain the result of the current output position.

[0105] Specifically, the pooling layer is used to reduce the data dimension while retaining important features and reducing computational complexity;

[0106] Specifically, the fully connected layer maps the features extracted by the convolutional layer to an output space of fixed dimensions for classification or regression;

[0107] Specifically, the output layer outputs the probability distribution of the two-dimensional graph classification corresponding to the single-phase latent grounding fault and the system asymmetry event.

[0108] It should be noted that the above step S300 not only improves the sensitivity and accuracy of fault detection, but also enhances the system's adaptability and generalization capabilities, and can maintain stable performance under different operating conditions. In addition, the above steps avoid the limitations of relying on preset thresholds in traditional methods, reduce the possibility of false alarms and missed alarms, and thus significantly improve the operational stability and safety of the power system, ensuring that the power grid can more reliably respond to various potential fault situations.

[0109] Example 2

[0110] Reference Figure 5 to Figure 13 Based on the previous embodiment, this embodiment provides an application example of a method for detecting latent grounding faults in a low-resistance grounding system, in order to verify and illustrate the technical effects adopted in this method.

[0111] This embodiment is built in the MATLAB / Simulink environment as follows Figure 5 The simulation model of single-phase latent ground fault in a 10kV neutral point low-resistance grounding system is shown. The neutral point grounding resistance of the system is set to 10 ohms. The system contains six feeders with lengths of 2km, 6km, 8km, 10km, 12km, and 20km respectively. The positive sequence parameters of the line are: R1 = 0.28Ω / km, X1 = 0.26mh / km, C1 = 0.38μF / km; the zero sequence parameters are: R0 = 2.8Ω / km, X0 = 1.11mh / km, C0 = 0.28μF / km. The single-phase latent grounding fault under different fault conditions (different fault lines (lines are: line L1, line L2, line L3, line L4, line L5, line L6), different fault time (fault time is 0.05s-0.07s), different transition resistance (transition resistance is 0.01 ohm-3000 ohm), with / without grounding arc) and events that may cause the neutral line zero-sequence current amplitude to increase (capacitor switching, unbalanced load switching) are set to obtain the neutral line zero-sequence current waveform. Some waveforms are shown as follows: When a single-phase arc grounding fault of 3000 ohms occurs on phase B of line L3, the neutral line zero-sequence current waveform is as follows: Figure 6 As shown, when load switching operation occurs in line L6, the zero-sequence current waveform of the neutral line is as follows: Figure 7 shown.

[0112] Set the neutral zero-sequence current amplitude setting value to 1.5A, and intercept the neutral zero-sequence current waveform within 0.005s before and 0.01s after the amplitude exceeds the limit. Set the sampling time to 0.001s, and convert the intercepted neutral zero-sequence current waveform near the amplitude exceeding the limit into a digital signal through the analog-to-digital conversion module.

[0113] The digital signal is converted into a two-dimensional graph through the Gram angle field. Some GASF and GADF graphs are shown in Figure 8 , Fig. 9 As shown in the figure, the two-dimensional graph corresponding to the waveform near the point where the amplitude of the neutral line zero-sequence current exceeds the set value under the influence of different events is labeled, the two-dimensional graph corresponding to the latent grounding fault is given label 1, and the two-dimensional graph corresponding to the system asymmetry event is given label 0.

[0114] The two-dimensional graphs with given labels are divided into training set and test set at a ratio of 4:1, and put into convolutional neural network for training. The convolutional neural network is used to classify the two-dimensional graphs corresponding to different events that cause the increase of the neutral line zero-sequence current amplitude to achieve the classification of latent grounding faults and system asymmetry, so as to realize the detection of single-phase latent grounding faults in low-resistance grounding systems. The training progress of the convolutional neural network accuracy, the training progress of the loss, the training results and the test results are shown in Figure 2. Fig.10 , Fig.11 , Fig.12 , Fig.13 shown.

[0115] It can be seen from the above simulation that the present invention establishes a simulation model of a single-phase latent grounding fault in a low-resistance grounding system to obtain the neutral line zero-sequence current waveform corresponding to events such as single-phase latent grounding faults and system asymmetric events that may cause the neutral line zero-sequence current amplitude to increase under different conditions; intercepts the waveform near the point where the neutral line zero-sequence current amplitude exceeds the set value, and converts it into a digital signal through a signal sampling module; converts the digital signal into a two-dimensional graph through a Gram angle field; after giving a label to the two-dimensional graph, it is used as training data to input into a convolutional neural network for training and testing, and the two-dimensional graphs corresponding to different events are classified to achieve effective distinction between single-phase latent grounding faults and system asymmetric events, which can not only accurately capture and identify high grounding resistance faults that are difficult to detect with traditional protection mechanisms, improve the speed and accuracy of fault diagnosis, but also reduce the false alarm rate, enhance the operational stability and safety of the power system, and provide strong technical support for timely troubleshooting, ensuring grid reliability, and maintaining public safety.

[0116] Example 3

[0117] This embodiment provides a latent ground fault detection system for a low-resistance grounding system, including a fault simulation module, a fault detection and waveform conversion module, and a fault classification module;

[0118] Specifically, the fault simulation module is used to establish a single-phase latent ground fault simulation model of a low-resistance grounding system and obtain a first current waveform;

[0119] Specifically, the fault detection and waveform conversion module is used to detect the existence of a latent ground fault and convert the intercepted second current waveform into a two-dimensional graph;

[0120] Specifically, the fault classification module is used to input the two-dimensional graph into the first neural network for training, so as to classify the two-dimensional graphs of different events and distinguish between single-phase latent grounding faults and system asymmetry events.

[0121] It should be noted that the technical solution of the system for detecting latent grounding faults in a low-resistance grounding system and the technical solution of the method for detecting latent grounding faults in a low-resistance grounding system mentioned above belong to the same concept. For details not described in detail in the technical solution of the system for detecting latent grounding faults in a low-resistance grounding system in this embodiment, please refer to the description of the technical solution of the method for detecting latent grounding faults in a low-resistance grounding system mentioned above.

[0122] 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.

[0123] This embodiment also provides an electronic device, which 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 implemented through WIFI, an operator network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a latent grounding fault detection method for a low-resistance grounding system 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 button, trackball or touchpad set on the computer device housing, or an external keyboard, touchpad or mouse, etc.

[0124] This embodiment also provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the method proposed in the above embodiment is implemented.

[0125] The storage medium proposed in this embodiment and the method proposed in the above embodiment belong to the same inventive concept. The technical details not fully described in this embodiment can be referred to the above embodiment, and this embodiment has the same beneficial effects as the above embodiment.

[0126] Through the above description of the implementation methods, the technicians in the relevant field can clearly understand that the present invention can be implemented by means of software and necessary general hardware, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ReadOnly, Memory, ROM), random access memory (RandomAccess Memory, RAM), flash memory (FLASH), hard disk or optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform the methods of various embodiments of the present invention.

[0127] 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.

[0128] 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.

[0129] 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.

[0130] 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.

[0131] 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 1The steps for the functions specified in one or more boxes.

[0132] 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.

[0133] 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 detecting latent ground faults in a low-resistance grounding system, characterized in that: include: Establish a single-phase latent ground fault simulation model for a low-resistance grounding system and obtain a first current waveform; Detecting the existence of a latent ground fault and converting the intercepted second current waveform into a two-dimensional graph; The two-dimensional graph is input into a first neural network for training to classify the two-dimensional graphs corresponding to different events, thereby distinguishing between single-phase latent grounding faults and system asymmetry events.

2. The method for detecting latent ground faults in a low-resistance grounding system according to claim 1, characterized in that: The detection of the latent ground fault includes: Setting a first current amplitude setting value; When the amplitude of the zero-sequence current of the system neutral line is greater than the first current amplitude setting value, it is determined that at least a single-phase latent grounding fault or a system asymmetry event occurs in the system.

3. The method for detecting latent ground faults in a low-resistance grounding system according to claim 2, characterized in that: The intercepted second current waveform is a zero-sequence current waveform within a first time threshold at the point where the amplitude exceeds the limit.

4. The method for detecting latent ground faults in a low-resistance grounding system according to claim 3, characterized in that: The converting the intercepted second current waveform into a two-dimensional graph comprises: Sampling the analog signal at a second time threshold to generate discrete sample points; quantizing the amplitude of the sampled discrete signal, converting it into a finite discrete value, and converting the discrete value into a binary form of a digital signal; Convert to 2D graphics via Gram's angle field.

5. The method for detecting latent ground faults in a low-resistance grounding system according to claim 4, characterized in that: The first neural network includes an input layer, a convolution layer, an activation function layer, a pooling layer, a fully connected layer and an output layer; The input layer receives a two-dimensional graph generated by Gram's angular field transformation, the convolution layer extracts features from the input data through a convolution kernel, the pooling layer is used to reduce the data dimension, the fully connected layer maps the features extracted by the convolution layer to an output space of a fixed dimension for classification or regression, and the output layer outputs a probability distribution of the two-dimensional graph classification corresponding to single-phase latent grounding faults and system asymmetry events.

6. The method for detecting latent ground faults in a low-resistance grounding system according to claim 1, characterized in that: The single-phase latent ground fault simulation model of the low-resistance grounding system includes: Neutral grounding resistance and feeder lengths of varying lengths; In the simulation model, the values ​​of the neutral point grounding resistance, feeders of different lengths, and the positive-sequence parameters and zero-sequence parameters of the line are set respectively.

7. The method for detecting latent ground faults in a low-resistance grounding system according to claim 4, characterized in that: The conversion of the Gram angle field into a two-dimensional graphic comprises: Collecting and preprocessing the amplitude data of the neutral line zero-sequence current, and representing the amplitude of the preprocessed neutral line zero-sequence current and its corresponding timestamp by the angle and radius of polar coordinates respectively; Divide the interval into several parts, associate the remaining points with the time series data by discarding 0, and make each point fall into a unit circle in polar coordinates; The GAF algorithm is used to reconstruct the neutral line zero-sequence current time series, and two different images are generated based on different formulas, namely the Gram angle sum field or the Gram angle difference field.

8. A system using the method for detecting latent ground faults in a low-resistance grounding system according to any one of claims 1 to 7, characterized in that: include: A fault simulation module, used to establish a single-phase latent ground fault simulation model of a low-resistance grounding system and obtain a first current waveform; A fault detection and waveform conversion module, used to detect the existence of a latent ground fault and convert the intercepted second current waveform into a two-dimensional graph; The fault classification module is used to input the two-dimensional graph into the first neural network for training, so as to classify the two-dimensional graphs of different events and distinguish between single-phase latent grounding faults and system asymmetric events.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method for detecting latent grounding faults in a low-resistance grounding system according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the steps of the method for detecting latent grounding faults in a low-resistance grounding system according to any one of claims 1 to 7.

Images