Line selection method for single-phase grounding fault in small current grounding system

Abstract

The present invention belongs to the technical field of fault detection in power systems, and proposes a line selection method for single-phase grounding faults in small current grounding systems, including collecting the fundamental component and the fifth harmonic component of the zero-sequence current of each line in the small current grounding system, and obtaining its inherent modal energy through empirical mode decomposition (EMD); then, these fault characteristics are used as input, BP neural network is used for fusion processing, and the line selection result is output. The line selection method for single-phase grounding faults in small current grounding systems proposed by the present invention has the advantages of high line selection accuracy, fast speed, good reliability, etc., and can be widely used in fault detection and positioning of power systems, providing a strong guarantee for the safe operation of power systems.

Description

Technical Field

[0001] The invention belongs to the technical field of power system fault detection and relates to a line selection method for a single-phase grounding fault in a small current grounding system. Background Art

[0002] In the power system, the low current grounding method has higher power supply reliability than the high current grounding method. When a single-phase grounding fault occurs in a low current grounding system, the line voltage between the three phases is still symmetrical. Since the neutral point is not directly connected to the earth, the fault current is very small, and the protection device does not need to trip and shut down the power immediately. It is generally allowed to operate for 1 to 2 hours. However, at this time, the non-fault relative voltage amplitude increases by √3 times, and the capacitive current to the ground also increases accordingly. If it is not handled in time, it may develop into a more serious fault for a long time, posing a greater threat to the safe operation of the entire power system. In order to detect and determine the fault grounding line and fault location in time, the traditional manual pulling method needs to disconnect the distribution lines one by one to determine the fault line, which is time-consuming, labor-intensive and inefficient. Frequent line start and stop will cause a certain impact on the power grid and power equipment, and affect the normal power supply of the line. With the advancement of technology, computer-based automatic line selection devices have begun to be used to automatically identify fault information by setting appropriate selection constraints, but there are still cases of misselection in actual applications. Summary of the invention

[0003] The present invention proposes a line selection method for a single-phase grounding fault in a small current grounding system, which solves the problem of low accuracy and slow speed in line selection and fault location when a single-phase grounding fault occurs in a small current grounding system, realizes convenient, fast and accurate fault location, reduces the heavy work of manual line selection, improves the work efficiency at the fault site, and reduces the power supply loss caused by the fault power outage time, with significant direct benefits. At the same time, the rapid line selection and location of the grounding fault enables the operation and maintenance personnel to quickly deal with related problems, reduce the operation time with the fault, and further reduce the probability of derivative accidents or damage to electrical equipment, which has extremely high direct and indirect benefits for ensuring the safe and stable operation of the distribution line and the stability of the power supply.

[0004] The technical solution of the present invention is achieved in this way:

[0005] A line selection method for a single-phase grounding fault in a small current grounding system comprises the following steps:

[0006] a) When a single-phase grounding fault occurs in the system, the zero-sequence current signal of each line is collected; the zero-sequence current signal is analyzed in the frequency domain by using fast Fourier transform to extract the fundamental component and the fifth harmonic component of the zero-sequence current;

[0007] b) performing empirical mode decomposition on the zero-sequence current to obtain its inherent mode energy;

[0008] c) taking the fundamental component, the fifth harmonic component and the inherent modal energy of the zero-sequence current as fault characteristics;

[0009] d) Using BP neural network, integrating the fault features, and outputting line selection results.

[0010] Preferably, the method further comprises step e): after selecting the faulty line, locating the fault point by using a BP neural network according to the changing rules of the fault characteristics.

[0011] Preferably, the acquisition in step a) is completed by a current transformer arranged in a small current grounding system.

[0012] Preferably, the empirical mode decomposition in step b) is to decompose the non-stationary, nonlinear zero-sequence current signal into a series of intrinsic mode functions, and calculate the energy of each IMF.

[0013] Preferably, the fault characteristics in step c) also include steady-state and transient information after the fault occurs.

[0014] Preferably, the fault point location in step e) is based on the influence of the distance of the fault point from the busbar on the fault characteristics, and the BP neural network is used to find out the change pattern of the fault characteristics to achieve the location of the fault point.

[0015] Preferably, the BP neural network learns and identifies the mapping relationship between fault characteristics and faulty lines through training sample data.

[0016] Preferably, the training sample data is obtained by simulating a line grounding fault in a low-current grounding system, and then collecting and processing the data.

[0017] Preferably, the following steps are also included: using MATLAB simulation software to build a simulation model of a small current grounding system; simulating a line grounding fault to obtain zero-sequence current waveforms of the fault line and the non-fault line; and verifying the effectiveness and limitations of the line selection method through simulation.

[0018] Preferably, the small current grounding system includes a neutral point ungrounded system and a neutral point grounded system via an arc suppression coil.

[0019] The working principle and beneficial effects of the present invention are:

[0020] 1. The present invention is based on the study of the transient and steady-state fault characteristics of single-phase grounding. It combines the BP neural network to normalize the steady-state information fault judgment criteria and the transient information fault judgment criteria, and then fuses them to use multiple information to jointly select the fault line and improve its accuracy; then the fault distance is used as the output of the neural network, and finally the single-phase grounding fault line selection and precise positioning of the small current grounding system based on the BP neural network is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0022] Figure 1 This is a simplified diagram of the low current grounding system of the present invention.

[0023] Figure 2 This is a flow chart of the line selection method of the present invention. DETAILED DESCRIPTION

[0024] The following will be combined with the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0025] like Figure 1 As shown in the figure, K1 disconnected is a simplified circuit diagram of the normal operation of the ungrounded system; K1 closed is a simplified circuit diagram of the normal operation of the neutral point grounded system through the arc suppression coil. When a single-phase grounding fault occurs in a small current grounding system, because the grounding current is very small, it is not necessary to immediately shut down the power, but it cannot operate with the fault for a long time, so it is of great significance to quickly and accurately select the faulty line. Before selecting the line, the electrical characteristics before and after the line fault must first be analyzed, and the fault line and the normal line must be distinguished based on the electrical changes before and after the fault. Then, the appropriate electrical quantity is selected as the fault characteristic, and a suitable line selection plan is formulated based on the selected fault characteristic information. Since there are almost no systems grounded by resistance in my country's small current grounding system, the present invention only considers two grounding methods: neutral point grounded through arc suppression coil and neutral point ungrounded.

[0026] like Figure 2 As shown, it is a flow chart of the line selection method of the present invention, which selects three fault characteristics of the fundamental component of the zero-sequence current of the line, the fifth harmonic component and the inherent modal energy of the zero-sequence current after EMD decomposition as line selection criteria, and uses the BP neural network to fuse the three information to output the line selection result.

[0027] (1) Steady-state information line selection

[0028] When a single-phase grounding fault occurs in the system, the fundamental amplitude of the zero-sequence current in the neutral point ungrounded system in steady state has obvious fault characteristics, and the fifth harmonic amplitude of the zero-sequence current in the neutral point arc suppression coil grounded system is less affected by the arc suppression coil, which can be used as a fault feature for line selection. Therefore, after collecting the zero-sequence current signal, frequency domain analysis is performed on it to extract the fundamental component and the fifth harmonic component, formulate fault criteria, and select the fault line. Fourier transform can decompose a signal into the superposition sum of several sinusoidal waveforms of different frequencies, and convert the time domain signal into a frequency domain signal. The signal processed by Fourier transform can reflect the statistical average result of the original signal content in the entire frequency domain.

[0029] (2) Transient information line selection

[0030] When a single-phase grounding fault occurs in the line, the zero-sequence current signal is mostly non-stationary. Therefore, the Hilbert-Huang transform is used to process the non-stationary signal, and the zero-sequence current signal is processed together with the Fourier transform to obtain the respective fault judgment criteria. If the line selection result based on the steady-state information is uncertain, the transient information is further used for line selection. The zero-sequence current signal is decomposed by EMD, the IMF is obtained and its energy is calculated, and the fault line is selected by using the energy difference between the fault line and the non-fault line.

[0031] (3) Simulation verification and optimization

[0032] A simulation model of a small current grounding system is built, and a line grounding fault is simulated to verify the effectiveness of the line selection method. At the same time, the factors affecting the magnitude of the zero-sequence current fundamental component and the fifth harmonic component under different grounding systems are analyzed to optimize the line selection method.

[0033] Through simulation analysis, the three fault characteristics of the line zero-sequence current, the fundamental component, the fifth harmonic component and the inherent modal energy of the zero-sequence current after EMD decomposition are selected as the line selection criteria, and the BP neural network is used to fuse the three information to output the line selection result.

[0034] (4) Operation interface and maintenance upgrade

[0035] Provide a user-friendly operation interface to facilitate operation and maintenance personnel to quickly master and use it. Regularly maintain and upgrade the line selection device to ensure the accuracy and reliability of the line selection method.

[0036] Example 1

[0037] 1. Signal acquisition

[0038] When a single-phase grounding fault occurs in the system, the zero-sequence current signal is collected through the zero-sequence current transformer installed on each line. The collected signal should have a high sampling rate and signal-to-noise ratio to ensure the accuracy of subsequent analysis.

[0039] 2. Frequency domain analysis

[0040] The collected zero-sequence current signal is analyzed in the frequency domain using Fast Fourier Transform (FFT) to extract the fundamental component and the fifth harmonic component of the zero-sequence current. The specific steps are as follows:

[0041] 3. Non-stationary signal processing

[0042] The zero-sequence current signal is processed using the Hilbert-Huang transform (HHT) to obtain a series of intrinsic mode functions (IMFs) and calculate their energy. The specific steps are as follows:

[0043] Hilbert-Huang transform: decomposes the zero-sequence current signal into multiple IMFs, each of which represents an inherent oscillation mode of the signal.

[0044] 4. Fault line selection

[0045] According to the fundamental component, fifth harmonic component and IMF energy value of zero-sequence current, the fault line is determined. The specific steps are as follows:

[0046] Threshold judgment: Set the thresholds of the fundamental component, the fifth harmonic component and the IMF energy. When the zero-sequence current signal of a line exceeds the threshold, the line is judged to be a fault line. Among them, setting the thresholds of the fundamental component, the fifth harmonic component and the IMF energy is one of the key steps in the fault line selection and location method. The determination of these thresholds usually needs to be combined with the characteristics of the actual system and experimental data, and is obtained through the following methods:

[0047] 1. Empirical method

[0048] The empirical method sets the threshold based on a large amount of actual operating data and the experience of engineers. This method is suitable for power systems that have accumulated rich operating experience. The specific steps are as follows:

[0049] Data collection: Collect zero-sequence current signal data of the system during normal operation and when a fault occurs.

[0050] Data analysis: Statistically analyze the distribution of the fundamental component, fifth harmonic component, and IMF energy of the zero-sequence current signal during normal operation to determine a reasonable upper limit.

[0051] Threshold setting: The threshold is set based on the maximum value during normal operation plus a certain safety margin (such as 10% to 20%).

[0052] 2. Statistical method

[0053] The statistical method is to determine the threshold through statistical analysis. The specific steps are as follows:

[0054] Data collection: Collect zero-sequence current signal data of the system during normal operation and when a fault occurs.

[0055] Data standardization: The collected data is standardized to make it conform to the normal distribution or other known distribution.

[0056] Threshold calculation: Based on the standardized data, calculate the mean and standard deviation of each characteristic value (fundamental component, fifth harmonic component and IMF energy). Usually the mean plus several times the standard deviation (such as 2σ or 3σ) is selected as the threshold.

[0057] 3. Machine Learning

[0058] The machine learning method automatically determines the threshold by training the machine learning model. The specific steps are as follows:

[0059] Data preparation: Collect the zero-sequence current signal data of the system during normal operation and when a fault occurs, and mark whether each piece of data is fault data.

[0060] Feature extraction: The fundamental component, fifth harmonic component and IMF energy of the zero-sequence current signal are extracted as features.

[0061] Model training: Use classification algorithms such as support vector machine (SVM), decision tree, random forest, etc. to train the data and determine the best classification boundary.

[0062] Threshold determination: The threshold is set according to the critical value of the classification boundary.

[0063] 4. Experimental verification method

[0064] The experimental verification method is to determine the threshold through actual experiments. The specific steps are as follows:

[0065] Experimental design: Different types of single-phase grounding faults are simulated in a laboratory environment, and the zero-sequence current signals of each line are recorded.

[0066] Data acquisition: Collect the fundamental component, fifth harmonic component and IMF energy of the zero-sequence current signal under different fault conditions.

[0067] Threshold setting: Based on experimental data, determine a threshold that can accurately identify the faulty line under various fault conditions.

[0068] 5. Comprehensive method

[0069] The comprehensive method combines the above methods to determine the threshold to improve the accuracy and robustness of the threshold. The specific steps are as follows:

[0070] Preliminary setting: First, the threshold is preliminarily set through empirical or statistical methods.

[0071] Verification and optimization: The initially set threshold is verified and optimized through experimental verification and machine learning methods to ultimately determine the optimal threshold.

[0072] Example

[0073] Assuming we determine the threshold by statistical method, the specific steps are as follows:

[0074] 1. Data collection: Collect 1000 sets of zero-sequence current signal data during normal operation.

[0075] 2. Data standardization: Standardize the data to make it conform to the normal distribution.

[0076] 3.Threshold calculation:

[0077] The average value of the fundamental component is 10A, the standard deviation is 2A, and the threshold is set to 10+3*2=16A.

[0078] The average value of the fifth harmonic component is 2A, the standard deviation is 0.5A, and the threshold is set to 2+3*0.5=3.5A.

[0079] The average value of IMF energy is 100 J, the standard deviation is 20 J, and the threshold is set to 100+3*20=160 J.

[0080] Through the above method, we can scientifically and reasonably set the thresholds of the fundamental component, fifth harmonic component and IMF energy, so as to effectively identify the faulty line.

[0081] Comprehensive evaluation: Combine multiple fault judgment criteria for comprehensive evaluation to improve the accuracy of fault line selection.

[0082] 5. Fault point location

[0083] Use BP neural network to find out the changing rules of fault criteria, and determine the location of the fault according to the acquired fault criteria. The specific steps are as follows:

[0084] BP neural network structure: The input layer includes three nodes, corresponding to the fundamental component, fifth harmonic component and IMF energy of the zero-sequence current; the output layer includes one node for outputting the location of the fault point.

[0085] Network training: Use known fault point data sets for supervised learning to optimize the parameters of the BP neural network and improve the accuracy and robustness of fault point location.

[0086] Real-time positioning: During system operation, the zero-sequence current signals of each line are collected in real time, and the trained BP neural network is used to automatically identify the fault line and determine the location of the fault point.

[0087] Experimental verification

[0088] 1. Simulation Model

[0089] Use MATLAB and other simulation software to build a simulation model of a small current grounding system, simulate the line grounding fault, and obtain the zero-sequence current waveform of the fault line and the non-fault line. The specific steps are as follows:

[0090] System modeling: Based on the actual grid parameters, a simulation model of the small current grounding system is established.

[0091] Fault setting: Set single-phase grounding faults at different locations and record the zero-sequence current signal of each line.

[0092] Data acquisition: Collect zero-sequence current signals through simulation software and save them as data files.

[0093] 2. Data Analysis

[0094] Perform frequency domain analysis and non-stationary signal processing on the collected zero-sequence current signal to extract the fundamental component, fifth harmonic component and IMF energy. The specific steps are as follows:

[0095] Frequency domain analysis: Use FFT to perform frequency domain analysis on the zero-sequence current signal and extract the fundamental component and the fifth harmonic component.

[0096] Non-stationary signal processing: Use HHT to process the zero-sequence current signal, obtain IMF and calculate its energy.

[0097] 3. Network training

[0098] Use the known fault point data set to train the BP neural network and optimize the network parameters. The specific steps are as follows:

[0099] Data preparation: The extracted fundamental component, fifth harmonic component and IMF energy are used as input and the fault point location is used as output to construct a training data set.

[0100] Network training: Use the training data set to supervise the BP neural network and adjust the network weights and biases until the predetermined error standard is reached.

[0101] Performance evaluation: The performance of the BP neural network is evaluated through the test data set, including indicators such as accuracy, recall, and F1 score.

[0102] 4. Experimental Results

[0103] Through simulation tests on different fault locations, the effectiveness and robustness of this method in fault line selection and location are verified. Experimental results show that this method can accurately identify the fault line and accurately locate the fault point under various fault conditions.

[0104] Application Examples

[0105] In a certain city distribution network, this method is used to select and locate single-phase grounding faults. The specific steps are as follows:

[0106] System configuration: Install zero-sequence current transformers on each line and connect them to the data acquisition system.

[0107] Data acquisition: Collect the zero-sequence current signal of each line in real time and transmit it to the central control system.

[0108] Fault detection: When an abnormal zero-sequence current is detected, the fault line selection and location program is started.

[0109] Fault handling: Based on the positioning results, the dispatcher takes timely measures to eliminate the fault and restore power supply.

[0110] In this actual fault, a single-phase grounding fault occurred in a line, resulting in a local power outage. Through this method, the central control system quickly identified the faulty line and accurately located the fault point. Based on the positioning results, the dispatcher sent maintenance personnel to the site to eliminate the fault in time and restore power supply. The whole process took less than 1 hour, significantly improving the reliability of power supply.

[0111] The above are only preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A line selection method for single-phase grounding fault in a small current grounding system, characterized in that: The following steps are involved: a) When a single-phase grounding fault occurs in the system, collect the zero-sequence current signal of each line; The zero-sequence current signal is analyzed in the frequency domain by using fast Fourier transform to extract the fundamental component and the fifth harmonic component of the zero-sequence current. b) performing empirical mode decomposition on the zero-sequence current to obtain its inherent mode energy; c) taking the fundamental component, the fifth harmonic component and the inherent modal energy of the zero-sequence current as fault characteristics; d) Using BP neural network, integrating the fault features, and outputting line selection results.

2. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The method further comprises step e): after selecting the faulty line, locating the fault point by using a BP neural network according to the changing rule of the fault characteristics.

3. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The acquisition in step a) is completed by a current transformer arranged in a small current grounding system.

4. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The empirical mode decomposition in step b) is to decompose the non-stationary, nonlinear zero-sequence current signal into a series of intrinsic mode functions and calculate the energy of each IMF.

5. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The fault characteristics in step c) also include steady-state and transient information after the fault occurs.

6. The line selection method for single-phase grounding fault in a small current grounding system according to claim 2 is characterized in that: The fault point location in step e) is based on the influence of the distance between the fault point and the busbar on the fault characteristics, and the BP neural network is used to find out the change law of the fault characteristics to locate the fault point.

7. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The BP neural network learns and identifies the mapping relationship between fault characteristics and faulty lines through training sample data.

8. The line selection method for single-phase grounding fault in a small current grounding system according to claim 7 is characterized in that: The training sample data is obtained by simulating a line grounding fault in a low-current grounding system, and then collecting and processing the data.

9. The line selection method for single-phase grounding fault in a small current grounding system according to claim 1 is characterized in that: The following steps are also included: Use MATLAB simulation software to build a simulation model of a small current grounding system; simulate and set a line grounding fault to obtain the zero-sequence current waveform of the fault line and the non-fault line; The effectiveness and limitations of the line selection method are verified through simulation.

10. A line selection method for a single-phase grounding fault in a small current grounding system according to any one of claims 1 to 9, characterized in that: The small current grounding system includes a neutral point ungrounded system and a neutral point grounded system via an arc suppression coil.

Images