A recording wave data preprocessing method for fault location under lightning stroke grounding fault condition

Abstract

The application provides a recording wave data preprocessing method for fault location under a lightning grounding fault condition, and the recording wave data preprocessing method is specifically as follows: an electromagnetic transient simulation model under a lightning condition is established, a fault recording waveform under a lightning grounding fault condition is obtained based on the electromagnetic transient simulation model under the lightning condition; the obtained fault recording waveform is subjected to noise reduction processing, the fault recording waveform after the noise reduction processing is intercepted to a preset length, a to-be-analyzed sample set is constructed by a moving sampling method based on the intercepted fault recording waveform; a fault starting moment and a fault ending moment are obtained according to the to-be-analyzed sample set, and a recording waveform before the fault occurs and a recording waveform after the fault occurs are obtained according to the fault starting moment and the fault ending moment. The application can perform noise reduction processing on the fault recording waveform, and improve the judgment accuracy of the fault starting moment and the fault ending moment, thereby improving the accuracy of subsequent fault location.

Description

Technical Field

[0001] This invention relates to the field of power transmission line lightning fault analysis technology, and in particular to a method for preprocessing waveform data for fault location under lightning grounding fault conditions. Background Technology

[0002] High-voltage transmission lines are the channels for transmitting electrical energy in power systems. However, due to the long transmission distances, they are easily affected by environmental and weather factors, leading to various types of faults. After a fault occurs, it is crucial to quickly and accurately locate the fault and restore power supply to ensure the safe and reliable operation of the power system. The waveform data recorded at both ends of the transmission line contains important information about the line and system faults. Mining and extracting useful information from this fault waveform data has significant theoretical and practical value for fault location.

[0003] Existing fault location algorithms based on fault waveform data often require preprocessing of the waveform data during computation. This preprocessing mainly includes two aspects: determining the fault start and end points from the on-site waveform data, and extracting the fundamental frequency data of various relevant electrical quantities of the faulty line. Accurate fault start and end times are crucial for fundamental frequency extraction.

[0004] Currently, the main technology for determining the fault initiation point in on-site waveform data mainly utilizes the sudden change current detection method, which can detect the fault start point in the waveform data based on changes in the sudden change current. However, the detection accuracy of the traditional sudden change current detection method is limited by many factors, such as wave refraction and reflection. Furthermore, under lightning strike conditions, because the rise time of lightning is on the order of microseconds, while the sampling frequency of the fault waveform is no more than 10kHz, it is difficult to accurately capture the fault initiation moment. Moreover, when a transmission line fault caused by lightning occurs, it introduces significant transient interference into the line at the moment of the fault. Combined with the inductive coupling between different phases, this results in a large amount of high-frequency transient noise in the waveform acquired during the fault. Since no noise reduction processing is performed on the waveform during fundamental wave extraction, it is impossible to accurately determine the fault initiation and termination moments, which may hinder the subsequent fault location accuracy. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and provide a preprocessing method for fault location recording data under lightning grounding fault conditions. This method obtains the fault recording waveform under lightning grounding fault conditions through model simulation, performs noise reduction processing on the fault recording waveform, and then determines the fault start and end times based on the noise-reduced fault recording waveform. This solves the problems of low accuracy in recording data acquisition and low accuracy in detecting the fault start and end times due to the lack of noise reduction processing in existing lightning fault data processing methods. It enables the acquisition of accurate fault start and end times, thereby improving the accuracy of fault location results under lightning grounding fault conditions.

[0006] The objective of this invention is achieved through the following technical solution:

[0007] A method for preprocessing waveform data for fault location under lightning-induced grounding fault conditions, comprising:

[0008] The voltage level information of the transmission line is obtained, and an electromagnetic transient simulation model under lightning strike conditions is established based on the voltage level of the transmission line. Based on the electromagnetic transient simulation model under lightning strike conditions, the fault waveform under lightning grounding fault conditions is obtained.

[0009] The acquired fault recording waveform is denoised, and the denoised fault recording waveform of a preset length is truncated. Based on the truncated fault recording waveform, a sample set to be analyzed is constructed using a moving sampling method.

[0010] Based on the constructed sample set to be analyzed, the fault start time and fault end time are obtained. Based on the fault start time and fault end time, the waveforms before and after the fault are obtained from the intercepted fault waveforms.

[0011] Furthermore, when performing noise reduction on the acquired fault recording waveform, the fault recording waveform is first decomposed using wavelet basis functions to obtain the wavelet decomposition coefficients corresponding to the fault recording waveform. Then, the decomposed fault recording waveform is thresholded using a wavelet threshold function. The fault recording waveform is reconstructed based on the wavelet coefficients after thresholding to obtain the noise-reduced fault recording waveform.

[0012] Furthermore, the expression for the wavelet coefficients after thresholding is as follows:

[0013]

[0014] in: C represents the wavelet coefficients after thresholding. i,j denoted as the wavelet decomposition coefficients of the noisy signal, T is the threshold, k is the adjustment coefficient, e is the natural constant, and sgn() is the sign function.

[0015] Furthermore, the sample set to be analyzed includes a fault initiation analysis sample set and a fault termination analysis sample set.

[0016] Furthermore, the specific process of constructing a fault initiation analysis sample set based on the intercepted fault waveform using the moving sampling method is as follows: obtain the maximum value in the intercepted fault waveform, and obtain the time t when the maximum value occurs. p And obtain the time t when the maximum value occurs. p The time t0 of the previous cycle is used to extract the time from time t0 to time t. p The fault waveform recorded within the time period is taken as the sample to be analyzed, S0. The moving sampling method is used to move the sampling window forward by one sampling interval Δt, and the time from t0-Δt to time t is truncated. p The fault waveform recorded within the time interval -Δt is taken as the sample S1 to be analyzed. The sampling window is then moved forward by one sampling interval Δt, and the time from t0-2Δt to t is extracted. p The fault waveform recorded within the time interval -2Δt is used as the sample to be analyzed, S2. The sampling window is then moved to obtain the fault initial analysis sample set S = {S0, S1, S2, ... S}. n}, and the sample S to be analyzed n The intercept time t0-nΔt and time t p The timing of the first three cycles is equal.

[0017] Furthermore, the specific process of constructing a fault termination analysis sample set based on the intercepted fault waveform using the moving sampling method is as follows: obtain the maximum value in the intercepted fault waveform, and obtain the time t when the maximum value occurs. p And obtain the time t when the maximum value occurs. p The time t1 of the next cycle, the intercept time t p The fault waveform recorded up to time t1 is used as the sample L0 to be analyzed. The moving sampling method is used to move the sampling window forward by one sampling interval Δt, and the waveform at time t is extracted. p The fault waveform recorded within the time interval from +Δt to t1+Δt is taken as the sample L1 to be analyzed. The sampling window is then moved forward by one sampling interval Δt, and the time t is extracted. p The fault waveform recorded during the time interval from +2Δt to t0+2Δt is used as the sample L2 to be analyzed. The sampling window is then moved to obtain the fault end analysis sample set L = {L0, L1, L2, ... L}. m}, and the sample L to be analyzed m The intercept time t p +mΔt and time t p The timing of the last five cycles is equal.

[0018] Furthermore, when obtaining the fault start time and fault end time based on the constructed sample set to be analyzed, the fault start time is obtained through the fault start analysis sample set S, and the fault end time is obtained through the fault end analysis sample set L.

[0019] Furthermore, the specific process of obtaining the fault start time through the fault start analysis sample set is as follows: add up all waveform values ​​of the fault recording waveform corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sample to be analyzed in the fault start analysis sample set S where the sum of waveform values ​​is 0, filter out the sample to be analyzed with the latest start time of the corresponding intercept time period, and take the start time of the filtered sample to be analyzed as the fault start time.

[0020] Furthermore, all waveform values ​​of the fault recording waveform corresponding to each sample to be analyzed in the fault end analysis sample set L are summed to obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault end analysis sample set L. The samples to be analyzed in the fault end analysis sample set L with a sum of waveform values ​​of 0 are obtained. The sample to be analyzed with the earliest start time of the corresponding intercept time period is selected, and the end time of the selected sample to be analyzed is taken as the fault end time.

[0021] The beneficial effects of this invention are:

[0022] This system can acquire corresponding fault waveforms by constructing an electromagnetic transient simulation model under lightning strike conditions and perform noise reduction on the fault waveforms, thereby improving the accuracy of subsequent fault start and end time determination. Furthermore, it can construct the sample set to be analyzed based on a moving sampling method and obtain accurate fault start and end times by utilizing the characteristics of electrical quantities operating under steady-state conditions through forward and backward summation. This provides accurate and complete waveform information before and after the fault for subsequent fault location algorithms, improving the accuracy of fault location results under lightning grounding fault conditions. Attached Figure Description

[0023] Figure 1 This is a schematic diagram of a process of the present invention;

[0024] Figure 2 This is a comparison diagram of a fault recording waveform and a fault recording waveform after noise reduction, according to an embodiment of the present invention.

[0025] Figure 3 This is a waveform diagram of the fault obtained according to an embodiment of the present invention;

[0026] Figure 4 This is a waveform diagram obtained after a fault, according to an embodiment of the present invention. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] Example:

[0029] A method for preprocessing waveform data for fault location under lightning-induced grounding fault conditions, such as... Figure 1 As shown, it includes:

[0030] The voltage level information of the transmission line is obtained, and an electromagnetic transient simulation model under lightning strike conditions is established based on the voltage level of the transmission line. Based on the electromagnetic transient simulation model under lightning strike conditions, the fault waveform under lightning grounding fault conditions is obtained.

[0031] The acquired fault recording waveform is denoised, and the denoised fault recording waveform of a preset length is truncated. Based on the truncated fault recording waveform, a sample set to be analyzed is constructed using a moving sampling method.

[0032] Based on the constructed sample set to be analyzed, the fault start time and fault end time are obtained. Based on the fault start time and fault end time, the waveforms before and after the fault are obtained from the intercepted fault waveforms.

[0033] The electromagnetic transient simulation model under lightning strike conditions adopts the Heidler model as the simulation model, and the electromagnetic transient simulation model specifically includes a transmission line frequency variation model, a tower and tower grounding resistance model, a lightning current model, and an insulator flashover model.

[0034] The frequency-changing model of transmission lines can be based on the actual parameters of the line, such as conductor parameters and line height, and adopts the frequency-changing line J.Marti model.

[0035] The lightning current model can simulate a lightning current signal with a waveform of 2.6 / 50μs.

[0036] When constructing the tower and tower grounding resistance model, the transmission line voltage level is first classified. When the transmission line is a 10kV or 35kV distribution network, a lumped inductance model is used as the tower and tower grounding resistance model, and the average inductance is taken as 0.5μH / m. When the line voltage level does not exceed 110kV and the wave process on the tower is not a concern, a single wave impedance model is used, and the wave impedance is taken as 150Ω. When the line voltage exceeds 220kV, a multi-wave impedance model is used.

[0037] When constructing the insulator flashover model, the 50% discharge voltage U of the insulator string is first calculated based on the actual model of the insulator string being simulated. 50％ Then, an insulator flashover model is established based on the corresponding voltage-controlled switch.

[0038] The discharge voltage U 50％ The calculation formula is:

[0039] U 50％ =533L x +132;

[0040] Where: L x This refers to the length of the insulator string.

[0041] The acquired fault waveforms include voltage and current signals.

[0042] When performing noise reduction on the acquired fault recording waveform, the fault recording waveform is first decomposed using wavelet basis functions to obtain the wavelet decomposition coefficients corresponding to the fault recording waveform. Then, the decomposed fault recording waveform is thresholded using wavelet threshold functions. The fault recording waveform is reconstructed based on the wavelet coefficients after thresholding to obtain the noise-reduced fault recording waveform.

[0043] Specifically, the fault recording waveform is first decomposed using the Haar wavelet basis function, and then thresholding is achieved using a wavelet threshold function based on the Logistic function.

[0044] The expression for the wavelet coefficients after threshold processing is as follows:

[0045]

[0046] in: C represents the wavelet coefficients after thresholding. i,j denoted as the wavelet decomposition coefficients of the noisy signal, T is the threshold, k is the adjustment coefficient, e is the natural constant, and sgn() is the sign function.

[0047] The larger k is, the faster the transition, and the steeper the transition region of the threshold function. In this embodiment, k is taken as 0.2.

[0048] When extracting the noise-reduced fault waveform of a preset length, the preset length is generally taken as 3-4 cycles before the fault occurs to 3-5 cycles after the fault occurs. Since the operating frequency of the transmission line is 50Hz, in this embodiment, the fault waveform is specifically extracted from 0.06s-0.08s before the fault occurs to 0.06-0.1s after the fault occurs, with the overall extracted fault waveform length being 0.1-0.2s.

[0049] The sample set to be analyzed includes a fault initiation analysis sample set and a fault termination analysis sample set.

[0050] The specific process of constructing a fault initiation analysis sample set based on the intercepted fault waveform using a moving sampling method is as follows: Obtain the maximum value in the intercepted fault waveform, and obtain the time t at which the maximum value occurs. p And obtain the time t when the maximum value occurs. p The time t0 of the previous cycle, specifically, since the normal operating frequency of the transmission line is 50Hz and the duration of one cycle is 0.02s, therefore, t0 = t p -0.02. The time interval from t0 to t1 is used. p The fault waveform recorded within a certain time period is taken as the sample to be analyzed, S0. Specifically, the time period for the sample to be analyzed, S0, is (t0, t...). p The moving sampling method is used, which moves the sampling window forward by one sampling interval Δt, and intercepts the time from t0-Δt to time t. p The fault waveform recorded within the time interval -Δt is taken as the sample to be analyzed, and the time interval of the sample to be analyzed, S1, is (t0-Δt, t p -Δt], move the sampling window forward by one sampling interval Δt, and extract the time from t0-2Δt to time t. p The fault waveform recorded within the time interval -2Δt is used as the sample to be analyzed, S2. The sampling window is then moved to obtain the fault initial analysis sample set S = {S0, S1, S2, ... S}. n}, and the sample S to be analyzed n The intercept time t0-nΔt and time t p The first three cycles are at the same time, i.e., t0 - nΔt = t p -0.06, and n must be rounded down.

[0051] The specific process of constructing a fault termination analysis sample set based on the intercepted fault waveform using the moving sampling method is as follows: Obtain the maximum value in the intercepted fault waveform, and obtain the time t when the maximum value occurs. p And obtain the time t when the maximum value occurs. p The time t1 of the next cycle, specifically, t1 = t p +0.02, intercept time t p The fault waveform recorded up to time t1 is taken as the sample L0 to be analyzed. Specifically, the time period of the sample L0 to be analyzed is (t p [t1], using a moving sampling method, the sampling window is moved backward by one sampling interval Δt, and the intercept time t is... p The fault waveform recorded during the time interval from +Δt to t1+Δt is taken as the sample L1 to be analyzed. The time interval of the sample L1 to be analyzed is (t p +Δt, t1+Δt], shift the sampling window back by one sampling interval Δt, and intercept time t. pThe fault waveform recorded during the time interval from +2Δt to t0+2Δt is used as the sample L2 to be analyzed. The sampling window is then moved to obtain the fault end analysis sample set L = {L0, L1, L2, ... L}. m}, and the sample L to be analyzed m The intercept time t p +mΔt and time t p The last five cycles have the same timing, i.e., t p +mΔt=t p +0.1, similarly, m needs to be rounded down.

[0052] When obtaining the fault start time and fault end time based on the constructed sample set to be analyzed, the fault start time is obtained through the fault start analysis sample set S, and the fault end time is obtained through the fault end analysis sample set L.

[0053] The specific process of obtaining the fault start time through the fault start analysis sample set is as follows: add up all waveform values ​​of the fault waveform corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sample to be analyzed in the fault start analysis sample set S where the sum of waveform values ​​is 0, filter out the sample to be analyzed with the latest end time of the corresponding intercepted time period, and take the end time of the filtered sample to be analyzed as the fault start time.

[0054] Add up all waveform values ​​of the fault recording waveform corresponding to each sample to be analyzed in the fault termination analysis sample set L to obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault termination analysis sample set L. Obtain the sample to be analyzed in the fault termination analysis sample set L where the sum of waveform values ​​is 0. Filter out the sample to be analyzed with the earliest start time of the corresponding intercept time period, and take the start time of the filtered sample to be analyzed as the fault termination time.

[0055] Due to the characteristics of periodic signals, the signal is symmetrically distributed within each period, so its sum should theoretically be 0. However, due to sampling accuracy issues, the obtained value will be close to 0 but not reach 0. Therefore, after obtaining the sum of waveform values, normalization is performed. The sum of waveform values ​​with an absolute value less than a preset value is considered 0, and in other cases, it is considered 1. Since the sample S0 to be analyzed includes the maximum waveform value under lightning strike fault, the sum of all waveform values ​​in sample S0 will be much greater than 0. Therefore, the starting time corresponding to the first sample to be analyzed with a waveform value sum of 0, i.e., the sample closest to sample S0, is the fault start time. Similarly, after sample L0, the ending time corresponding to the first sample to be analyzed with a waveform value sum of 0, i.e., the sample closest to sample L0, is the fault end time.

[0056] After filtering out the fault start and end times by summing the waveform values, the fault recording waveform obtained can be a standard trigonometric function waveform, which belongs to the electrical quantity under steady-state operation and can be used as the data basis for subsequent fundamental wave extraction and fault location algorithm.

[0057] This embodiment takes a 500kV transmission line as an example. This transmission line is a double-ended 500kV power supply. One side has a voltage of 523.1∠-10.8°kV, with a system equivalent positive-sequence impedance of 1.05 + j43.18Ω and a zero-sequence impedance of j29.09Ω. The other side has a voltage of 522.7∠-16.2°kV, with a system equivalent positive-sequence impedance of 1.06 + j44.92Ω and a zero-sequence impedance of j37.47Ω. A set of 75Mvar high-voltage reactors is installed on each power supply side, with a high-voltage rated voltage of 525kV, a reactance value of 2016Ω, and a neutral point reactance of 1099Ω. The transmission line maintains a double-circuit configuration on the same tower throughout its entire length, spanning 65km. A frequency-varying model of the transmission line is constructed using the J. Marti model. The lightning current is simulated using a 2.6 / 50μs waveform, and the overall simulation is achieved using the Heidler model. For the tower and tower grounding resistance models, a multi-wave impedance model is employed based on the voltage level.

[0058] The fault waveform was obtained through simulation and then denoised. A comparison of the fault waveform and the denoised fault waveform is shown in the figure below. Figure 2 As shown.

[0059] based on Figure 2 It can be seen that the maximum value occurs at time t. p =0.0638s, therefore, t0 = 0.0438s, t1 = 0.0838s. Furthermore, in this embodiment, the sampling interval Δt is set to 0.0001s, therefore n is 200 and m is 1000.

[0060] By summing the waveform values ​​of each sample to be analyzed in the fault initiation analysis sample set S, we can obtain h = {h0, h1, h2, ..., h...} n}, and h0 corresponds to S0, h1 corresponds to S1, h2 corresponds to S2, ..., h n Corresponding to S n ,based on Figure 2 From this, we know that h0 = -675450.5, h1 = -991022.8, h2 = -1281759.0, h3 = -1530256.0, ... until S. 16 The corresponding waveform value sum h 16 =0.0117, while S 15 The corresponding waveform value sum h 15 =-28883.8, then h17 =0.0017,h 18 =0.0118,h 19 =0.0118…, indicating that S 16 The subsequent waveforms represent normal operating conditions, therefore S can be obtained. 16 The end time t p -16Δt=0.0638-16×Δt=0.0622s, meaning the fault initiation time is 0.0622s.

[0061] Similarly, by summing the waveform values ​​of each sample to be analyzed in the fault termination analysis sample set L, we can obtain d = {d0, d1, d2, ..., d...} m}, and d0 corresponds to L0, d1 corresponds to L1, d2 corresponds to L2, ..., d m Corresponding to L m ,based on Figure 2 From this, we know that d0 = -599447.0, d1 = -599696.5, d2 = -599946.1, d3 = -600195.6, ... up to L. 280 The corresponding waveform value sum d 280 =0.0102, while L 281 The corresponding waveform value sum d 281 =0.0104, d 282 =0.0107, d 283 =0.0103, d 284 =0.0108…, indicating L 280 The subsequent waveforms represent normal operating conditions, therefore L can be obtained. 280 The starting time t p +280Δt=0.0638+280×0.0001=0.0918s, that is, the fault end time is 0.0918s.

[0062] Combination Figure 2 The waveform changed significantly at 0.0622s and stopped changing significantly after 0.0918s, returning to a normal periodic waveform. This also verifies the accuracy of the fault start and end times.

[0063] After obtaining the fault start and end times, the waveforms before and after the fault are obtained, and their waveform diagrams are as follows. Figure 3 , Figure 4 As shown.

[0064] The embodiments described above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any way. Other variations and modifications are possible without departing from the technical solutions described in the claims.

Claims

1. A method for preprocessing waveform data for fault location under lightning-induced grounding fault conditions, characterized in that, include: The voltage level information of the transmission line is obtained, and an electromagnetic transient simulation model under lightning strike conditions is established based on the voltage level of the transmission line. Based on the electromagnetic transient simulation model under lightning strike conditions, the fault waveform under lightning grounding fault conditions is obtained. The acquired fault recording waveform is denoised, and the denoised fault recording waveform of a preset length is truncated. Based on the truncated fault recording waveform, a sample set to be analyzed is constructed using a moving sampling method. The fault start time and fault end time are obtained from the constructed sample set to be analyzed. Based on the fault start time and fault end time, the waveform before the fault occurred and the waveform after the fault occurred are obtained from the intercepted fault waveform. The sample set to be analyzed includes a fault initiation analysis sample set and a fault termination analysis sample set; The specific process of constructing a fault initiation analysis sample set based on the intercepted fault waveform using the moving sampling method is as follows: Obtain the maximum value in the intercepted fault waveform, and obtain the time when the maximum value occurs. And obtain the time when the maximum value occurs. The moment of the previous cycle Capture time At that time The fault waveform recorded within the time period is taken as the sample S0 to be analyzed. The moving sampling method is used to move the sampling window forward by one sampling interval. Capture time At that time The fault waveform recorded within the time period is taken as the sample to be analyzed, S1, and the sampling window is moved forward by one sampling interval. Capture time At that time The fault waveform recorded within the time period is used as the sample to be analyzed, S2. The sampling window is then moved to obtain the fault initial analysis sample set S = {S0, S1, S2, ... S...}. n }, and the sample S to be analyzed n The interception time With time The first three cycles are at the same time; The specific process of constructing a fault termination analysis sample set based on the intercepted fault waveform using the moving sampling method is as follows: Obtain the maximum value in the intercepted fault waveform, and obtain the time when the maximum value occurs. And obtain the time when the maximum value occurs. The moment of the next cycle Capture time At that time The fault waveform recorded within the time period is taken as the sample L0 to be analyzed. A moving sampling method is used to shift the sampling window backward by one sampling interval. Capture time At that time The fault waveform recorded within the time period is used as the sample L1 to be analyzed, and the sampling window is then moved forward by one sampling interval. Capture time At that time The fault waveform recorded within the time period is used as the sample to be analyzed, L2. The sampling window is then moved to obtain the fault end analysis sample set L={L0, L1, L2, ...L... m }, and the sample L to be analyzed m The interception time With time The timing of the last five cycles is equal; When obtaining the fault start time and fault end time based on the constructed sample set to be analyzed, the fault start time is obtained through the fault start analysis sample set S, and the fault end time is obtained through the fault end analysis sample set L. The specific process of obtaining the fault start time through the fault start analysis sample set is as follows: add up all waveform values ​​of the fault waveform corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault start analysis sample set S, obtain the sample to be analyzed in the fault start analysis sample set S where the sum of waveform values ​​is 0, filter out the sample to be analyzed with the latest end time of the corresponding intercepted time period, and take the end time of the filtered sample to be analyzed as the fault start time.

2. The method for preprocessing waveform data for fault location under lightning grounding fault conditions according to claim 1, characterized in that, When performing noise reduction on the acquired fault recording waveform, the fault recording waveform is first decomposed using wavelet basis functions to obtain the wavelet decomposition coefficients corresponding to the fault recording waveform. Then, the decomposed fault recording waveform is thresholded using wavelet threshold functions. The fault recording waveform is reconstructed based on the wavelet coefficients after thresholding to obtain the noise-reduced fault recording waveform.

3. The method for preprocessing waveform data for fault location under lightning grounding fault conditions according to claim 2, characterized in that, The expression for the wavelet coefficients after threshold processing is as follows: ； in: These are the wavelet coefficients after thresholding. denoted as the wavelet decomposition coefficients of the noisy signal, T is the threshold, k is the adjustment coefficient, e is the natural constant, and sgn() is the sign function.

4. The method for preprocessing waveform data for fault location under lightning grounding fault conditions according to claim 1, characterized in that, Add up all waveform values ​​of the fault recording waveform corresponding to each sample to be analyzed in the fault termination analysis sample set L to obtain the sum of waveform values ​​corresponding to each sample to be analyzed in the fault termination analysis sample set L. Obtain the sample to be analyzed in the fault termination analysis sample set L where the sum of waveform values ​​is 0. Filter out the sample to be analyzed with the earliest start time of the corresponding intercept time period, and take the start time of the filtered sample to be analyzed as the fault termination time.

Images