Calibration method for initial traveling wave head of high-resistance grounding fault of power distribution network

Abstract

The invention discloses a power distribution network high-resistance grounding fault initial traveling wave head calibration method, and relates to the field of power distribution system protection and fault positioning. The method comprises the steps of obtaining a fault voltage signal of a measurement node in a power distribution network, and calculating a line mode component of the fault voltage signal; extracting a fault transient component from the line mode component; calculating an integral deviation waveform of a fault transient component to select a fault sudden change time window; a multi-scale time window differentiator is constructed, fault initial traveling wave arrival time is accurately calibrated in a fault sudden change time window, and accurate time data is provided for power distribution network fault traveling wave positioning, so that accurate fault positioning is realized.

Description

Technical Field

[0001] This invention relates to the field of power system protection and fault detection, and in particular to a method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks. Background Technology

[0002] The distribution network serves as the link between the transmission network and power users. Accurate fault location in the distribution network is fundamental to improving the reliability of demand-side power supply and is a crucial link in ensuring the safe and stable operation of the entire power system. High-resistivity grounding faults occur frequently in the distribution network, and their fault characteristics are extremely weak. Attenuation is more severe on cable lines, and the wavefront abrupt change is more gradual, greatly increasing the difficulty of wavefront time calibration. Accurate acquisition of the traveling wavefront time of distribution network faults is of great significance for precise fault location.

[0003] In recent years, experts and scholars have mostly used time-frequency decomposition algorithms to denoise signals, and then employed the Teager energy operator to detect fault signal abrupt changes. One method uses the wavelet-Teager energy operator to calibrate the arrival time of the transient voltage modulus signal at the measurement point. Other scholars perform ESMD decomposition on the fault signal, using the Teager energy operator to enhance the instantaneous energy of the first-level modal components and extract fault abrupt changes. Alternatively, they first use CEEMD to filter out low-frequency components in the signal, and then propose an improved Teager energy operator to perform differential operations on the signal to calibrate the traveling wavefront.

[0004] However, while the aforementioned method effectively enhances the instantaneous energy change characteristics of fault signals by introducing the Teager energy operator, extreme conditions such as high-impedance grounding result in extremely weak fault signals and severe wavefront attenuation during the propagation of traveling waves in cable lines. After noise is added, the fault characteristics are almost completely obscured. Denoising the signal inevitably alters the original characteristics of the fault abrupt change point, and the Teager energy operator is also affected by random noise, leading to the failure of abrupt change point time calibration. Summary of the Invention

[0005] The purpose of this invention is to provide a method for calibrating the initial traveling wave front of a high-resistivity ground fault in a distribution network, which solves the problem of detecting the initial traveling wave front of a high-resistivity ground fault in a distribution network in the prior art.

[0006] To achieve the above objectives, embodiments of the present invention provide a method for initial traveling wave front calibration of a high-resistivity grounding fault in a distribution network, comprising the following steps:

[0007] Step 1: Obtain the fault voltage signal at the measuring point. The fault voltage signal at the measuring point refers to the voltage signal at the measuring point after a high-resistance grounding fault occurs in the distribution network. Perform a Kelvin-Bell phase-mode transformation on the fault voltage signal at the measuring point to obtain the line-mode component and the zero-mode component, and select the line-mode component for subsequent fault abrupt change point detection. The fault voltage signal includes the three-phase voltage sampling signals of points A, B, and C. This step mainly converts the three-phase signals of A, B, and C into line-mode and zero-mode components. Subsequent fault abrupt change point detection is all based on processing the line-mode component.

[0008] Step 2: Reconstruct the power frequency signal of the linear mode component using a swarm intelligence optimization algorithm, extract the power frequency component from the linear mode component, and then subtract the reconstructed power frequency component from the linear mode component to obtain the fault transient component U. f ;

[0009] Step 3: Perform integration (i.e. summation) on the transient components of the fault, calculate the integral deviation (also known as cumulative deviation) waveform, select a subsequence, and obtain the fault abrupt change time window;

[0010] Step 4: Construct a multi-scale time window differential, calibrate the fault mutation point within the fault mutation time window, and obtain the initial traveling wave arrival time, thus realizing the initial traveling wave front calibration.

[0011] In step 1, the process of performing Kelvin phase-mode transformation on the three-phase signals at measurement points A, B, and C is as follows:

[0012] According to the formula Calculate the line-mode and zero-mode components of the fault voltage signal, where U A U B U C The fault voltage at the measuring point consists of the A, B, and C phase signals, U. α U β U0 and U0 represent the α-linear mode component, β-mode component, and zero-mode component, respectively.

[0013] In step 2, using the amplitude, frequency, and initial phase angle parameters of the trigonometric function as solutions, and aiming to minimize the error between the reconstructed power frequency signal and the original signal, a signal reconstruction parameter optimization model is established: the objective function is... in,

[0014] The signal is the power frequency reconstructed signal; T is the length of the time window used for signal reconstruction; X is the signal error.

[0015] A, f, and δ0 are the parameters to be optimized, representing the amplitude, frequency, and initial phase angle of the reconstructed power frequency signal, respectively. The reconstructed power frequency component is obtained through the particle swarm optimization algorithm in the swarm intelligence optimization algorithm. The fault transient component U is obtained by subtracting the reconstructed power frequency component from the α linear mode component. fSwarm intelligence optimization algorithms belong to heuristic algorithms, which is a broad category encompassing many specific algorithms. This invention employs the particle swarm optimization algorithm.

[0016] In step 3:

[0017] For the fault transient component U f After summing, the integral deviation E(t) is obtained: This step is equivalent to integrating a continuous signal: using integration operations. Calculate the integral deviation waveform, where U f E(t) represents the fault transient component; E(t) represents the integral deviation.

[0018] In step 4:

[0019] Traversing all time points in the integral deviation E(t), time t0 is selected, where t0 satisfies the condition that the absolute value of the difference between the average value of the integral deviation in the time series before t0 and the average value in the time series at t0 is maximized. This yields the initial fault window: [t0-T w / 2,t0+T w / 2],T w Let be the size of the time window. All data within this time window of E(t) are extracted as a subsequence e(t) of E(t). One t0 corresponds to one subsequence.

[0020] In step 4:

[0021] Based on the waveform characteristics of the integral deviation subsequence, slope fluctuation index and endpoint offset index are constructed;

[0022] A multi-scale time window differencer is constructed using the slope fluctuation index and the endpoint offset index.

[0023] The constructed slope fluctuation index includes:

[0024] use Standardize the data for e(t), where This is the standardized subsequence of the integral bias.

[0025] e(i) is the i-th element in e(t), where i is a subscript or index value.

[0026] Calculate the slope of the straight line connecting each point on the integral deviation subsequence to the point (t0, x(t0)).

[0027] Calculate the slope fluctuation index of the integral deviation subsequence before time t using the slope value. Slope fluctuation index after time t of the integral deviation subsequence

[0028] The constructed endpoint offset index includes:

[0029] The offset index is constructed using the offset values ​​between the start / end points of the integral deviation subsequence and x(t0), including the first offset index: Compared with the second offset index:

[0030] The construction of the multi-scale time window difference includes:

[0031] Constructing a multi-scale time window differencer P(t) is the multi-scale time window difference value (where the numerator is the difference between the start and end point offset index and the slope difference, and the denominator is the sum of the slope fluctuation index).

[0032] The step of calibrating the fault mutation point within the fault mutation time window and obtaining the initial traveling wave arrival time includes:

[0033] The integral bias subsequence is processed using a multi-scale time window differencer, and the result is obtained through a discriminant. The fault mutation point t is determined, that is, t that satisfies the formula is the mutation point. The initial traveling wave arrival time t is obtained, where T is the fault mutation time window length, which is taken as 200; λ is taken as 20.

[0034] Beneficial effects:

[0035] This invention proposes a method for calibrating the initial traveling wave front of a high-resistivity grounding fault in a distribution network. The method includes acquiring the fault voltage signal from a measurement node in the distribution network and calculating the line-mode component of the fault voltage signal; extracting the fault transient component from the line-mode component; calculating the integral deviation waveform of the fault transient component to select a fault abrupt change time window; constructing a multi-scale time window differential to accurately calibrate the arrival time of the initial traveling wave within the fault abrupt change time window, providing accurate time data for fault traveling wave localization in the distribution network, thereby achieving precise fault location. Using this method, the subtle characteristics of high-resistivity grounding faults can be highlighted, and the initial traveling wave front time can be accurately calibrated, providing a preliminary data feature basis for fault protection and fault location. Attached Figure Description

[0036] To more clearly illustrate the technical solutions of this invention, the accompanying drawings used in some embodiments of this invention will be briefly described below. Obviously, the drawings described below are only drawings of some embodiments of this invention, and those skilled in the art can obtain other drawings based on these drawings. Furthermore, the drawings described below can be regarded as schematic diagrams and are not intended to limit the actual size of the product, the actual flow of the method, the actual timing of the signals, etc. involved in the embodiments of this invention.

[0037] Figure 1 This is a flowchart of a method for initial traveling wave front calibration of a high-resistivity grounding fault in a distribution network, according to some embodiments.

[0038] Figure 2 The original waveform and the noise-enhanced waveform according to some embodiments;

[0039] Figure 3 A component waveform diagram reconstructed from a power frequency signal according to some embodiments;

[0040] Figure 4 The following is a waveform diagram of the transient components of a fault according to some embodiments;

[0041] Figure 5 Here is a waveform diagram of the integral deviation according to some embodiments;

[0042] Figure 6 The results of multi-scale time window differencer calculations based on some embodiments;

[0043] Figure 7 The waveform diagram at the measurement point is shown when the transition resistance is 0.1Ω according to some embodiments;

[0044] Figure 8 The results of fault wavefront calibration are shown in some embodiments when the transition resistance is 0.1Ω.

[0045] Figure 9 The waveform diagram at the measurement point is shown according to some embodiments when the transition resistance is 4kΩ;

[0046] Figure 10 The results show the fault wavefront calibration when the transition resistance is 4kΩ according to some embodiments. Detailed Implementation

[0047] The technical solutions in some embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments provided by the present invention are within the scope of protection of the present invention.

[0048] Embodiments of the present invention provide a method for initial traveling wave front calibration of a high-resistivity grounding fault in a distribution network. The overall flowchart of the method is shown below. Figure 1 As shown, firstly, the fault voltage signal of the measurement node in the distribution network is acquired, the line-mode component of the fault voltage signal is calculated, and the fault transient component is extracted from the line-mode component. Then, the integral deviation waveform of the fault transient component is calculated to select the fault abrupt change time window. Finally, a multi-scale time window differential is constructed to accurately calibrate the arrival time of the initial traveling wave of the fault within the fault abrupt change time window. Figure 2The waveform is the voltage waveform at the measuring point after a single-phase ground fault occurs in a 10kV distribution network. The grounding resistance is 4kΩ, and the sampling frequency is 10MHz. The original waveform shows that the fault signal amplitude is very small compared to the power frequency signal; after being superimposed with 40dB of noise, the fault characteristics are almost completely obscured.

[0049] A Kallenbehl phase-mode transformation was performed on the three-phase signals at measurement points A, B, and C, and the linear mode component was selected to detect fault abrupt changes. Using the amplitude, frequency, and initial phase angle parameters of the cosine function as solutions, and aiming to minimize the error between the reconstructed power frequency signal and the original signal, an optimization model for signal reconstruction parameters was established.

[0050]

[0051] In the formula: U α These are the original signal line-mode components; The signal is the power frequency reconstructed signal; T is the length of the time window used for signal reconstruction; X is the signal error; A, f, and δ0 are the parameters to be optimized, representing the amplitude, frequency, and initial phase angle of the reconstructed cosine signal, respectively.

[0052] This case study utilizes the PSO algorithm to solve the signal reconstruction parameter optimization model, and then uses the optimized parameters to reconstruct the power frequency signal over the entire time window. The results are as follows: Figure 3 As shown. Subtracting the power frequency reconstructed signal from the linear mode component yields the fault transient component, as shown. Figure 4 As shown in the figure, the extracted fault transient components exhibit significant differences in waveform characteristics before and after the fault abrupt change point.

[0053] To accurately determine the arrival time of the fault traveling wave, the transient component signal of the fault is further processed using equation (3) to obtain the signal integral deviation waveform, as shown below. Figure 5 As shown, the signal integral deviation waveform can reflect the cumulative deviation of the fault transient component from zero: before the arrival of the fault traveling wave, the signal only contains noise, which fluctuates around zero and the cumulative deviation is small; after the arrival of the fault traveling wave, the signal is superimposed with the transient component, and the cumulative deviation gradually increases.

[0054]

[0055] Iterate through all times from the integral deviation E(t), and select time t0 according to equation (4). t0 satisfies the condition that the absolute value of the difference between the average value of the integral deviation in the time series before t0 and the average value in the time series at t0 is the largest, thus obtaining the initial fault window: [t0-T w / 2,t0+T w / 2],T w Given the time window size, extract the subsequence e(t) of the integral deviation E(t) based on the fault time window;

[0056]

[0057] In equation (4), L is the length of the integral deviation signal.

[0058] Extract a subsequence e from the cumulative deviation curve E of the signal, centered at time t0:

[0059]

[0060] In the formula: T w +1 represents the length of the subsequence time window.

[0061] Construct slope fluctuation index H respectively v The local fluctuations and overall offset of the integral deviation subsequence are characterized by the use of (t) and the endpoint offset index ΔO(t):

[0062]

[0063]

[0064] In equations (6)-(8): k(i) is the slope of the straight line connecting each point on the signal curve to the point (t, E(t)); H v1 H represents the slope deviation before time t. v2 This represents the slope deviation after time t.

[0065]

[0066] In equations (9)-(10): ΔO1(t) and ΔO2(t) represent the offset values ​​between the start and end points of the signal and E(t), respectively.

[0067] A multi-scale time window difference P(t) is constructed using the signal slope, slope fluctuation index, and endpoint offset index:

[0068]

[0069] Calculate P(t) for the integral deviation subsequence; the waveform is as follows: Figure 6 As shown. Based on the following discriminant, the calibration time is 18.9 μs, the actual time is 19 μs, and the error is 0.1 μs.

[0070]

[0071] In the formula: T is the length of the fault change window, which is taken as 200; λ is taken as 20.

[0072] Build a power distribution network model in PSCAD and set a single-phase ground fault with an initial phase angle of 90°. Set the transition resistances to 0.1Ω and 4kΩ respectively. Add 50, 40, 30, and 20dB noise to the side point waveforms respectively.

[0073] When the actual fault arrival time is 19μs and the transition resistance is 0.1Ω, the measured node waveform is as follows: Figure 7 As shown, the waveform after the multi-scale time window differencer operation is as follows: Figure 8 As shown. When the transition resistance is 4kΩ, the measured node waveform is as follows. Figure 9 As shown, the waveform after the multi-scale time window differencer operation is as follows: Figure 10 As shown, the method proposed in this invention can accurately calibrate the initial traveling wave front of a high-resistivity grounding fault. Even when the fault characteristics are extremely weak, the calibration error of the proposed method is within 1 μs, demonstrating high detection accuracy.

[0074] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for initial traveling wave front calibration of a high-resistivity grounding fault in a distribution network, characterized in that, include: Step 1: Obtain the fault voltage signal at the measuring point. The fault voltage signal at the measuring point refers to the voltage signal at the measuring point after a high-resistance grounding fault occurs in the distribution network. Perform a Kelvin-Bell phase-mode transformation on the fault voltage signal at the measuring point to obtain the line-mode component and the zero-mode component, and select the line-mode component for subsequent fault change point detection. The fault voltage signal includes the three-phase voltage sampling signals at points A, B, and C. Step 2: Reconstruct the power frequency signal of the linear mode component using a swarm intelligence optimization algorithm, extract the power frequency component from the linear mode component, and then subtract the reconstructed power frequency component from the linear mode component to obtain the fault transient component U. f ; Step 3: Perform integration on the transient components of the fault, calculate the integral deviation waveform, select a subsequence, and obtain the fault abrupt change time window; Step 4: Construct a multi-scale time window differential, calibrate the fault mutation point within the fault mutation time window, and obtain the initial traveling wave arrival time, thus realizing the initial traveling wave front calibration.

2. The method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks according to claim 1, characterized in that, In step 1, the process of performing Kelvin phase-mode transformation on the three-phase signals at measurement points A, B, and C is as follows: According to the formula Calculate the line-mode and zero-mode components of the fault voltage signal, where U A U B U C The fault voltage at the measuring point consists of the A, B, and C phase signals, U. α U β U0 and U0 represent the α-linear mode component, β-mode component, and zero-mode component, respectively.

3. The method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks according to claim 1, characterized in that, In step 2, using the amplitude, frequency, and initial phase angle parameters of the trigonometric function as solutions, and aiming to minimize the error between the reconstructed power frequency signal and the original signal, a signal reconstruction parameter optimization model is established: the objective function is... in, Let be the power frequency reconstructed signal; T be the length of the time window used for signal reconstruction; X be the signal error; A, f, and δ0 be the parameters to be optimized, representing the amplitude, frequency, and initial phase angle of the reconstructed power frequency signal, respectively; the reconstructed power frequency component is obtained through the particle swarm optimization algorithm in the swarm intelligence optimization algorithm, and the fault transient component U is obtained by subtracting the reconstructed power frequency component from the α linear mode component. f .

4. The method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks according to claim 3, characterized in that, In step 3: For the fault transient component U f After summing, the integral deviation E(t) is obtained:

5. The method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks according to claim 1, characterized in that, In step 4: Traversing all time points in the integral deviation E(t), time t0 is selected, where t0 satisfies the condition that the absolute value of the difference between the average value of the integral deviation in the time series before t0 and the average value in the time series at t0 is maximized. This yields the initial fault window: [t0-T w / 2,t0+T w / 2],T w Let e ​​be the size of the time window. Extract all data from this time window of E(t) as a subsequence e(t) of E(t).

6. The method for initial traveling wave front calibration of high-resistivity grounding fault in distribution network according to claim 1, characterized in that, In step 4: Based on the waveform characteristics of the integral deviation subsequence, slope fluctuation index and endpoint offset index are constructed; A multi-scale time window differencer is constructed using the slope fluctuation index and the endpoint offset index.

7. The method for initial traveling wave front calibration of high-resistivity grounding faults in distribution networks according to claim 6, characterized in that, The constructed slope fluctuation index includes: use Standardize the data for e(t), where The standardized subsequence of the integral bias; Calculate the slope of the straight line connecting each point on the integral deviation subsequence to the point (t0, x(t0)). Calculate the slope fluctuation index of the integral deviation subsequence before time t using the slope value. Slope fluctuation index after time t of the integral deviation subsequence 8. The method for initial traveling wave front calibration of high-resistivity grounding fault in distribution network according to claim 7, characterized in that, The constructed endpoint offset index includes: The offset index is constructed using the offset values ​​between the start / end points of the integral deviation subsequence and x(t0), including the first offset index: Compared with the second offset index:

9. The method for initial traveling wave front calibration of high-resistivity grounding fault in distribution network according to claim 8, characterized in that, The construction of the multi-scale time window difference includes: Constructing a multi-scale time window differencer P(t) is the multi-scale time window difference value.

10. The method for initial traveling wave front calibration of high-resistivity grounding fault in distribution network according to claim 1, characterized in that, The step of calibrating the fault mutation point within the fault mutation time window and obtaining the initial traveling wave arrival time includes: The integral bias subsequence is processed using a multi-scale time window differencer, and the result is obtained through a discriminant. The fault mutation point t is determined, that is, t that satisfies the formula is the mutation point. The initial traveling wave arrival time t is obtained, where T is the fault mutation time window length, which is taken as 200; λ is taken as 20.

Images